chapter 6 quantitative research designs experimental quasi experimental and descriptive

• High School
• You don't have any recent items yet.
• You don't have any courses yet.
• You don't have any books yet.
• You don't have any Studylists yet.
• Information

Nutrition Research chapter 6

Stratford high - stratford-wi.

Recommended for you

Students also viewed.

• Felipe Gomes - Conclusion question period 4
• GSTForm - Hhh
• MD 22-23 MISA PT webready
• [NHÓM 1] [CHỦ ĐỀ 17] HOÁ DƯỢC
• Chasing corals summary
• 01 03 earths early etmosphere

Related documents

• Daily Lesson Plan Template oscar avah mihn
• Yomil Morales - sssss
• Dance Pass - idk
• '94 Fleece Jogger Old Navy
• 600535 0008714682 E2W23ES
• Copy of Chemistry of Life Phenomenon

Preview text

Quantitative research designs:, experimental, quasi-experimental,, and descriptive, chapter outline, ▶ introduction, ▶ experimental study designs, ▶ quasi-experimental designs, ▶ descriptive quantitative designs, ▶ additional types of designs, ▶ researcher interview: intervention research,, dr. leslie cunningham-sabo, learning outcomes, ▶ discuss five considerations when planning a, research design., ▶ explain the three essential components of, experimental designs, and compare and, contrast the following experimental designs:, randomized controlled trials, crossover, factorial,, and solomon four group designs., ▶ discuss the advantages and disadvantages of, various experimental designs., ▶ compare and contrast the nonequivalent, control group and interrupted time series, various quasi-experimental designs., ▶ compare and contrast the descriptive cross-, sectional, repeated cross-sectional, comparative,, and descriptive correlational designs., various descriptive designs., ▶ read a research study and identify the design, used and analyze study results., ▶ distinguish between secondary data analysis, and secondary research..

© Chad Baker/Getty Images

INTRODUCTION

Designing a research study requires making a number of decisions on the steps you, will take to answer your research question(s). like an architect, you need to prepare a, blueprint for your project. if you have ever met with an architect before, you know that, the process usually starts with a lot of questions. research design is no different. the, following questions address a number of key design features that must be considered., 1. what is the research question will there be an intervention testing the effects of an, intervention is the hallmark of experimental and quasi-experimental research., if there is an intervention with human participants, the researcher will assign, participants to be exposed to the independent variable, such as a modified diet or, nutrient supplement, or be part of the control group. experimental and quasi-, experimental designs are used to test a hypothesis., 2. instead of an intervention, will researchers observe study participants and take measure-, ments for example, researchers might observe a group over a longer period, of time to see if exposure to certain factors (such as a diet high in fruits and, vegetables) affects their risk of disease. this type of design is called a cohort study, design. it is commonly used in the field of epidemiology, a discipline within, public health that looks at the rates of health-related states (such as disease) in dif-, ferent groups of people and why they occur, and then looks at how this informa-, tion can be used to control health problems. study designs used in epidemiology, are discussed in chapter 7., 3. what are the variables what comparisons are going to be made between or within, groups comparisons are needed to examine relationships between the indepen-, dent and the dependent variable., 4. when and how often will data be collected or measurements taken many experimental, studies measure the dependent variable at least before and after the intervention., weight loss studies, for example, often take measurements for a year or more, to see whether participants kept the weight off. data may be collected at just, one point in time, such as in a cross-sectional study, or more frequently. in a, longitudinal study, participants are observed and measurements are taken over, a long period of time. longitudinal studies either go forward in time (prospec-, tive) or backward in time (retrospective)., 5. what will the setting be for the study the setting could be a hospital, community, center, or other location. some studies use multiple sites., 6. in an intervention study with at least two groups, will the participants be randomly assigned, to a group true experimental research involves random assignment to groups, so participants each have an equal chance of receiving any of the treatments, (including no treatment) under study. quasi-experimental research does not, have randomization of participants to groups., 7. in a human intervention study, will participants, researchers, and staff be blinded from, knowing to which group a participant was assigned blinding helps to prevent or, minimize sources of bias, such as expectation bias. expectation bias is when, researchers’ expectations of what they believe the study results should be get in, the way of accurately taking measurements and reporting results., 8. what controls will be put in place to reduce the influence of extraneous variables extra-, neous variables are factors outside of the variables being studied that might, influence the outcome of a study and cause incorrect conclusions. a good quan-, titative design identifies and rules out as many of these competing explanations, as possible., a good research design helps you answer the research question while effectively reducing, threats to design validity., quantitative research designs are often used to look at causal relationships, but they, can also be used to look at associations or relationship between variables. quantitative, research studies can be placed into one of five categories, although some categories do vary, 156 chapter 6: quantitative research designs: experimental, quasi-experimental, and descriptive, table 6 statistics that look at differences and statistics that look at associations, statistics that look at differences, statistic purpose number of groups, measurement, level of dependent, independent samples, t to test the difference between, the means of two independent, 2 interval/ratio, paired samples t-test, (or dependent t-test), the means from two paired, groups (such as before-and-, after observations on the same, one-way analysis of, variance (anova), f to test the difference among, means of more than two, independent groups for one, independent variable (with more, than one level)., more than two, interval/ratio, two-way analysis of, means for two independent, variables, of which each can have, multiple levels., repeated measures, anova (one-way, within-subjects), three or more means in the, same group over time. (extended, design of dependent samples, one group interval/ratio, chi-square χ 2 to analyze nominal and ordinal, data to find differences between, two or more groups nominal/ordinal, statistics that look at associations, statistic purpose, pearson product-, moment correlation, r to measure the strength and direction of the relationship, between two variables., spearman rank-, order correlation, ρ to measure the strength and direction of the relationship, between two variables. (nonparametric version of, pearson product-moment correlation), ordinal, interval, or, linear regression to predict the value of a dependent variable and, measure the size of the effect of the independent, variable on a dependent variable while controlling for, covariates., logistic regression same as linear regression; used when dependent value, binary/dichotomous, 158 chapter 6: quantitative research designs: experimental, quasi-experimental, and descriptive, formed. randomization also forms the basis for statistical testing. to randomize, participants, researchers first generate random numbers (using either computer, software or a random number table) and use them to assign each participant to a, group. this is referred to as simple randomization., in a randomized block design, the participants are first split into homoge-, neous groups, or blocks, before being randomly assigned to either the treatment, or control group within the block. blocks are used to decrease the variability of, the sample and to control the effects of a characteristic that could influence the, outcome, such as sex, age, weight, or severity of disease. each block generally, needs at least 20 participants (grove, burns, & gray, 2013). for example, a study, of the effect of omega-3 fatty acids on adults with a history of heart disease may, be randomized by age (see figure 6)., random assignment means that the groups will be comparable, and differences between the groups at, the end of the experiment can be deduced as being a result of the intervention. without randomiza-, tion of participants into groups, a study is considered to be quasi-experimental., figure 6. 1 simple randomization and randomized block.

####### Simple randomization

####### Random assignment

####### All subjects

####### n = 120

####### Treatment

####### group

####### n = 60

####### Control

####### n = 28

####### n = 27

####### n = 32

####### n = 33

####### Subjects

####### < 65 years

####### old

####### n = 55

####### 65 years

####### old and up

####### n = 65

####### Randomized block

Randomization into the treatment or control group is essential to an experimental study, but it

Is not essential that the participants be randomly selected from a target population before being, randomly assigned to a group. most randomized controlled trials do not use random sampling to pick who is, in the study, but all do use random assignment to groups., when reading research, you will come across two similar terms—control group and, comparison group—and probably will wonder what the difference is. the terms are often, used interchangeably, but there is a difference. technically, a control group is chosen, by random sampling of the target population, and a comparison group is chosen using, experimental study designs 159, group. this type of practice in a study may bias the estimate of the treatment effect by, 30 to 40% (moher et al., 1998; schulz, chalmers, hayes, & altman, 1995)., pretest data, sometimes called baseline data, is useful along with demographic data, to evaluate whether randomization really produced equivalent groups. normally, this, comparison of baseline data is discussed right at the beginning of the results section and, is displayed in one or more tables., randomization facilitates blinding, another feature of rcts. the cochrane hand-, book defines blinding as follows:, in general, blinding (sometimes called masking) refers to the process by which study, participants, health providers and investigators, including people assessing out-, comes, are kept unaware of intervention allocations after inclusion of participants, into the study. blinding may reduce the risk that knowledge of which intervention, was received, rather than the intervention itself, affects outcomes and assessments of, outcomes. (higgins & green, 2011, box 8.11), appropriate blinding can help reduce sources of bias such as performance bias (system-, atic differences between groups in the care provided)., in a single-blind study, participants are not told whether they are in the experi-, mental or the control group. in a double-blind study, two groups have been blinded—, normally the participants and one or all of these groups: health care providers, data, collectors, data analysts, and the researchers themselves (who may have a number of, roles in the study such as data collector). because the term double-blind lacks a standard, definition, you will need to read the study to see who was really blinded., when someone involved in a research study is responsible for assessing partici-, pant outcomes and knows which intervention a participant received, that person could, bias how the outcome was measured (usually reporting greater effects in the treatment, group). lack of blinding in rcts has been shown to inflate the effect of the interven-, tion by 9% (pildal et al., 2007)., blinding is possible in some, but not all, studies. in a drug study, both groups can, take pills as long as they look and taste the same. in a diet study where the intervention, group follows a modified diet and the control group eats their usual diet, blinding is not, possible. also keep in mind that even when a researcher can use blinding, it is not a, simple procedure and it does not always work perfectly., attrition is another concern when conducting rcts. if dropouts and noncompliant, participants are excluded from the data, it can cause a number of problems: it reduces, sample size and may disrupt the balance of characteristics in each group, thereby biasing, the results. for example, if more participants in the experimental group drop out than, from the control group (perhaps the intervention caused some of this), this creates an, imbalance. a technique called intention-to-treat analysis is used to prevent biases, due to participant attrition. intention-to-treat is the principle that all participants are used, in the statistical analysis, regardless of whether they dropped out, did not receive all the, treatments, or did not comply with the treatments. intention-to-treat has both support-, ers and detractors and advantages and disadvantages, but it is generally preferred. some, researchers modify intention-to-treat by, let us say, excluding certain participants., when reading results of an rct, examine how the study handles the following to, determine possible sources of bias:, 1. power calculation to determine sample size., 2. randomization and allocation to groups., 3. type(s) of blinding used (if any)., 4. follow-up of participants and intention-to-treat analysis (if used)., experimental study designs 161, 5. data collection., 6. precise, complete results., in a clinical trial, researchers test new treatments, drugs, or medical devices with, human participants to assess efficacy and safety. clinical trials are intervention studies, that often use an rct design. sometimes you also hear the term controlled clinical, trials (cct). controlled clinical trials do use a control group but may not assign par-, ticipants to the intervention or control group in a strictly random manner, making them, quasi-experimental studies., the fda requires (and regulates) clinical trials before a new drug, medical products, such as vaccines, or medical device is sold in the united states. clinical trials are often, done in stages or phases, each designed to answer a different research question (national, institutes of health, 2008)., phase i: researchers test a new drug or treatment in a small group of people for the, first time to evaluate its safety, determine a safe dosage range, and identify side, phase ii: the drug or treatment is given to a larger group of people to see if it is, effective and to further evaluate its safety., phase iii: the drug or treatment is given to large groups of people to confirm its, effectiveness, monitor side effects, compare it to commonly used treatments, and, collect information that will allow the drug or treatment to be used safely., phase iv: studies are done after the drug or treatment has been approved and mar-, keted to gather information on the drug’s effect in various populations and any, side effects associated with long-term use., clinical trials may be carried out in multiple locations simultaneously. there are a num-, ber of advantages to such multicenter studies: increased sample size, a more representa-, tive sample, more cost-effective, and increased generalizability of results. coordination, and communication in multicenter studies is, of course, more challenging than a single-, center trial., the u. national institutes of health maintains clinicaltrials, a web site that is a registry of, more than 200,000 public and privately supported clinical studies of human participants in the, united states and around the world. the web site also contains results and is used by patients, researchers,, students, and study record managers., most rcts use a pretest-posttest design, as shown in table 6. the pretest and, posttest are designated respectively as “o 1 ” and “o 2 .” think of “o” as an observation, in which data are collected and measurements taken. many rcts are simply variations, of these designs; some may use multiple experimental groups, perhaps receiving treat-, ment that varies by intensity, frequency, or duration. sometimes there may be more, than one comparison group, such as one comparison group that receives no treatment, and another comparison group that receives a placebo (as in a drug study)., table 6 randomized controlled trial design with pretest-posttest, random assignment experimental group o 1 treatment o 2, control group o 1 o 2, 162 chapter 6: quantitative research designs: experimental, quasi-experimental, and descriptive, table 6 summary of selected outcome measurements over time using linear effects model, intervention ( n = 93), mean (95% ci), placebo ( n = 99) mean, mean difference (95% ci), between treatment groups, systolic bp, baseline 121 (118 to 124) 118 (115 to 121) 2 (–1 to 7), 6 months 126 (122 to 129) 123 (120 to 127) 2 (–2 to 7), 12 months 125 (122 to 128) 123 (120 to 126) 1 (–2 to 6), diastolic bp, baseline 77 (75 to 79) 76 (74 to 78) 0 (–2 to 4), 6 months 79 (77 to 82) 79 (76 to 81) 0 (–2 to 3), 12 months 77 (75 to 79) 76 (74 to 78) 0 (–2 to 3), se glucose (mmol/l), baseline 5 (4 to 5) 4 (4 to 5) 0 (–0 to 0), 6 months 5 (4 to 5) 5 (4 to 5) 0 (–0 to 0), 12 months 5 (4 to 5) 5 (4 to 5) –0 (–0 to 0), se insulin (mu/l), baseline 13 (10 to 17) 11 (7 to 14) 2 (–2 to 7), 6 months 13 (11 to 15) 12 (10 to 14) 0 (–1 to 3), 12 months 13 (11 to 16) 12 (10 to 15) 1 (–2 to 4), homa-insulin resistance, baseline 3 (2 to 5) 2 (1 to 3) 1 (–0 to 3), 6 months 3 (2 to 3) 2 (2 to 3) 0 (–0 to 1), 12 months 3 (2 to 3) 2 (2 to 3) 0 (–0 to 1), tg (mmol/l), baseline 1 (1 to 1) 1 (1 to 1) –0 (–0 to 0), 6 months 1 (1 to 1) 1 (1 to 1) 0 (0 to 0)*, 12 months 1 (1 to 1) 1 (1 to 1) 0 (–0 to 0), hdl-c (mmol/l), baseline 1 (1 to 1) 1 (1 to 1) 0 (–0 to 0), 6 months 1 (1 to 1) 1 (1 to 1) –0 (–0 to 0).

• Significant at P < 0. Data from “Effect of Vitamin D Supplementation on Cardiometabolic Risks and Health-Related Quality of Life among Urban Premenopausal Women in a Tropical Country – A Randomized Controlled Trial,” by M. Ramly, M. Ming, K. Chinna, S. Suboh, and R. Pendek, 2014, PLOS ONE, 9, e110476, Table 2.

difference is 1. In this table, if these means were statistically significant, the authors would place an

Asterisk next to the mean difference., there was no significant effect of vitamin d on blood pressure, insulin resistance, triglycerides, or hdl, between the groups (all p > 0) except for the effect on triglycerides at 6 months. the results from the, 164 chapter 6: quantitative research designs: experimental, quasi-experimental, and descriptive, health-related quality-of-life questionnaire showed small but significant improvement, in vitality (mean difference: 5; 95% ci: 0 to 9) and mental component, score (mean difference: 2; 95% ci: 0 to 5) in the intervention group com-, pared to the placebo group., the consort 2010 checklist (see appendix c) describes information to include, when reporting a randomized trial. (consort stands for consolidated standards of, reporting trials.) the consort group, a panel of experts, developed this checklist, to increase the transparency of rcts and to reveal when there are deficiencies (schulz,, altman, & moher, 2010)., advantages of rcts include good internal validity and the use of powerful statistical, tests, such as analysis of variance (anova), to analyze data. rcts often can be used in, meta-analysis. rcts are the most appropriate research design to answer research ques-, tions on treatment or therapy. as for disadvantages, randomized experiments do tend, to be costly and time-consuming. rcts often suffer from noncompliance and dropouts, (sometimes due to side effects), and participants may respond differently because they, know they are being observed and assessed (known as the hawthorne effect). as in, many research studies, their external validity (ability to generalize results) may have, limitations. threats to external validity are minimized when a broadly representative, sample is used and the setting is not too controlled., crossover designs, in a crossover design, each participant acts as a member of both the experimental and, the control group. studies designed to compare two different groups of participants are, referred to as between-groups design. crossover designs are within-groups design, because the researcher is making comparisons within the same participants., the most common crossover design is the two-period, two-treatment design. par-, ticipants are randomly assigned to receive either the treatment in period 1 and the con-, trol in period 2, or the reverse. for example, in a drug study, one participant initially, received the active drug and then later received the placebo. to avoid carryover effects, (when exposure to a treatment affects outcomes in a later period), researchers build in, a period of time—called a washout period—between treatments for the effect of the, treatment to disappear., crossover studies include a design feature known as repeated measures. when, you see “repeated measures” in a study, it means multiple, repeated measurements are, being taken, not just a pretest and posttest., troup et al. (2015) used a crossover design to study the effects of black tea intake, (specifically the flavonoids in tea) on blood cholesterol levels on participants with mild, hypercholesterolemia. participants were block-randomized by sex to drink either 5, cups/day of black tea or 5 cups/day of the placebo for the first 4 weeks. the placebo was, a caffeinated beverage that looked and tasted like tea but contained no flavonoids. after, the 3-week washout period, participants switched assignments. figure 6 shows the, design for this study., during the treatment periods, participants were provided and consumed a low-, flavonoid diet. during the run-in periods (13 days each), participants drank the tea-like, placebo. participants were allowed to add sugar, but not milk, to either beverage (milk, reduces the antioxidant capacity of tea)., the study did not show that black tea significantly changed the lipid profile of the, participants. as in the study just mentioned, the researchers here also looked at the mean, difference of an outcome, such as ldl-c. none had a p-value below 0, the level of, significance., experimental study designs 165, receiving both the zinc and multivitamin supplements), you can use three groups and, a control group as shown in figure 6. locks et al. found that daily zinc supplements, starting in infancy had small but significant improvement in weight for age., this was an example of a simple factorial design. you could, for example, design a, diet/exercise study with four diets and three exercise programs. that would result in 12, combinations of diet/exercise, leading to a 4 × 3 factorial design., solomon four-group design, the solomon four-group design (table 6) is a combination of the pretest-posttest, design and the posttest only design. in this design, participants are randomly assigned to, one of two intervention groups or one of two control groups. both intervention groups, receive the same intervention; the only difference is that one of these groups receives the, pretest, the other does not. likewise, only one of the control groups receives the pretest., posttest measures are collected on all four groups to assess the effect of the independent, variable. some researchers modify this design and use just one control group, which, receives both the pretest and the posttest., atlantis, salmon, and bauman (2008) used a solomon four-group design to explore, the effects of television advertisements (independent variable) promoting physical activ-, ity on children’s preferences for physical or sedentary activities (dependent variables)., the children were randomized to one of two treatment groups or one of two control, groups. the treatment groups watched a television show with standard advertisements, and also advertisements promoting more physical activity instead of sedentary activity., the control groups watched the same show but without the advertisements promoting, physical activity. one experimental group and one control group were assessed before, and after watching the television show for their choices, preferences, and ratings of physi-, cal and sedentary activities. the other groups were only assessed after watching the tele-, vision show. the study did not show any significant differences between groups., this type of design is useful when a researcher thinks the outcomes could be biased, by exposure to the pretest. in general, the solomon four-group design is considered a, very rigorous design that strengthens both internal and external validity. as you can.

####### Zinc supplement +

####### Multivitamin

####### Multivitamin supplement

####### Supplement

####### (Independent

####### Variable)

####### Zinc Supplement

####### (Independent Variable)

####### Zinc supplement

####### only

####### supplement only

####### Control group

####### (placebo)

FIGURE 6. 3 Example of a 2 × 2 Factorial Design

Table 6 solomon four-group design, random assignment experimental group 1 o 1 treatment o 2, control group 1 o 1 o 2, experimental group 2 treatment o 2, control group 2 o 2, experimental study designs 167, imagine, this design is more time consuming for researchers and also requires a large, sample due to the four groups., quasi-experimental designs, quasi-experimental designs have an intervention and manipulation of the independent, variable, but they lack a key feature of experimental studies—randomization. because, we are unsure if the groups are truly equivalent, quasi-experimental designs are ranked, lower than experimental studies as sources of evidence., two of the most popular quasi-experimental designs are nonequivalent control, group and time series designs. be cautious when you see a quasi-experimental study, that does not have a control group. without a control group, a study has little, if any,, external or internal validity., nonequivalent control group designs, in most cases, a nonequivalent control group design is similar to the classic experi-, mental design except that participants are not randomly assigned to groups. often, researchers use natural groups or assign participants to groups using a nonrandom, method. sampling is still going on in terms of choosing the study’s participants. it is, just that participants do not have the same chance of being in either the experimental or, control group, and as a result, the groups are not necessarily equivalent., some researchers match participants at the group level based on demographic or, other possible confounding variables. the more similar the groups are, the closer the, design approximates an experimental study. researchers confirm whether two groups, are comparable (especially on the dependent variable) at baseline by collecting and ana-, lyzing pertinent data, but that may not include all baseline differences in active variables., table 6 shows a nonequivalent control group design with a pretest and posttest., there are a number of variations on this design, such as posttest only with a control, group (sometimes a pretest is not possible or would flaw the results) or pretest and post-, test with two comparison treatments and a routine care comparison group., mcaleese and rankin (2007) carried out a nonequivalent control group study to, “determine whether adolescents who participated in a garden-based nutrition interven-, tion would increase their fruit and vegetable consumption more than those participating, in a nutrition education intervention without any garden activities (mcaleese & rankin,, 2007, p. 662). this study appears in appendix a., 1. participants: a convenience sample of 99 sixth-grade students in three different, schools were the participants. two schools had the experimental groups and one, school had the control group., 2. measurements: all students took pretests (three 24-hour food recalls) and posttests, (three 24-hour food recalls)., 3. intervention: both experimental groups participated in a 12-week nutrition edu-, cation curriculum, “nutrition in the garden.” experimental school 2 also, participated in gardening activities, maintaining and harvesting a garden with, vegetables, herbs, and strawberries. the control group received no intervention., table 6 nonequivalent control group pretest/posttest design, experimental group o 1 treatment o 2, 168 chapter 6: quantitative research designs: experimental, quasi-experimental, and descriptive, table 6 abstract of a study using a simple interrupted time series design, objectives we examined changes in meal selection by patrons of university food-service, operations when nutrition labels were provided at the point of selection., methods we used a quasi-experimental, single-group, interrupted time-series, design to examine daily sales before, during, and after provision of point-, of-selection nutrition labels. piecewise linear regression was employed to, examine changes in the average energy content of entrees and a paired t, test was used to detect differences in sales across the periods., results the average energy content of entrees purchased by patrons dropped, immediately when nutrition labels were made available at point of, selection and increased gradually when nutrition information was removed., there was no significant change in number of entrees sold or in revenues, between the two periods., conclusions use of nutrition labels reduced the average energy content of entrees, purchased without reducing overall sales. these results provide support for, strengthening the nutrition labeling policy in food-service operations..

Reproduced from “Improving Patrons’ Meal Selections Through the Use of Point-of-Selection Nutrition Labels,” by Y. H. Chu, E. A. Frongillo, S. J. Jones, & G. L. Kaye, 2009, American Journal of Public Health, 99 , p. 2001. Reprinted with permission.

FIGURE 6. 4 Average Energy Content of Entrées Sold Per Day in a Food-Service Operation

Before, during, and after provision of nutrition information at point of selection: columbus,, ohio, october 25–december 8, 2004..

Reproduced from “Improving Patrons’ Meal Selections Through the Use of Point-of-Selection Nutrition Labels,” by Y. H. Chu, E. A. Frongillo, S. J. Jones, & G. L. Kaye, 2009, American Journal of Public Health, 99 , p. 2003. Reprinted with permission.

more for the history effect. Table 6 shows the interrupted time series design with

Control group. some researchers also call this a multiple time series design. basically a, multiple time series design has more than one group., interrupted time series designs are flexible and can be used in a number of situations., this type of design is especially useful in the evaluation of community interventions, when rcts are impractical and too expensive, and you want to focus on measuring, changes in behaviors and outcomes over time., table 6 contains an abstract of a study using a simple interrupted time series design, (chu, frongillo, jones, & kaye, 2009). figure 6 displays the results of the study..

####### 620

####### 625

Pretreatment 1Pretreatment 3Pretreatment 5Pretreatment 7Pretreatment 9Pretreatment 11Pretreatment 13Treatment 1Treatment 3Treatment 5Treatment 7Treatment 9Treatment 11Treatment 13Posttreatment 1Posttreatment 3Posttreatment 5Posttreatment 7Posttreatment 9Posttreatment 11Posttreatment 13

####### Days of Study

####### 630

####### 635

####### 640

####### 645

####### 650

####### 655

####### 660

####### Average Kcal Per Entrée Sold

170 Chapter 6: Quantitative Research Designs: Experimental, Quasi-Experimental, and Descriptive

Descriptive quantitative designs, descriptive designs collect information about variables without changing the environ-, ment or manipulating any variables, so they do not look at possible cause and effect., they are different from observational designs in that they do not include comparison, groups. according to grove, burns, and gray (2013), descriptive designs “may be used, to develop theory, identify problems with current practice, justify current practice, make, judgments, or determine what others in similar situations are doing” (p. 215)., descriptive designs range from cross-sectional surveys (at one or multiple points, in time) to comparative designs (comparing two groups) to correlations (relationships, between two variables). you can think of many descriptive designs as creating a snap-, shot. we now take a look at three common descriptive designs., descriptive cross-sectional and repeated, cross-sectional design, in a cross-sectional study, data is collected at one point in time. a purely descriptive, cross-sectional study provides basic information about prevalence (number of existing, cases of a disease or health condition in a population) and distribution, as you can see in, these examples..

• Using data from the National Health and Nutrition Examination Survey (2009–

2012), researchers reported that men consumed an average of 14 cups of water

Per day, and women consumed 11 cups of water per day (rosinger & herrick,, 2016). this indicates that americans seem to be taking in adequate fluids..

• You hear frequent media reports about how American adults and children are

managing their weight. Using data from the National Health and Nutrition

Examination survey (2013–2014), a group of researchers announced that 16%, of children and adolescents (ages 2–19 years) are overweight and 17% are obese, (skinner, perrin, & skelton, 2016)..

• If you ever wondered how many people really use menu labeling at fast-food or

chain restaurants, researchers found that of adults who noticed nutrition labeling

At fast-food or chain restaurants, 25% reported frequent use of the information,, 32% reported moderate use, and 43% reported they never used it (lee-kwan,, pan, maynard, mcguire, & park, 2016). the data came, again, from a national, survey: the behavioral risk factor surveillance system. in this study, data from, 17 states was used., a repeated cross-sectional study generally collects the same data at multiple points, in time, and usually includes both descriptive and inferential statistics (to look at the, when reading a study with an interrupted time series design, you want to take a good look at the, chart (such as figure 6) that shows the trend of the measurements before, during, and after the, intervention. the graph is split into the three segments. the average kcalories per entrée sold was high before, the intervention, and then went down during the intervention. it is also interesting that once the intervention, stopped, the average kcalories slowly moved back to the preintervention numbers. looking at a graph such, as this can give you a quick mental picture of what happened during a study., descriptive quantitative designs 171, table 6 secular trends from 1999 to 2010 in adolescent meal patterns by sociodemographic, characteristics: minneapolis-st. paul, mn, project eat (eating and activity in teens), breakfast frequency, (mean days/wk), lunch frequency, characteristic 1999 a (n) 2010 (n) 1999 a 2010 p valueb 1999 a 2010 p valueb, total sample 2,598 2,540 3 4 <0 5 5 <0., male 1,181 1,175 4 4 0 5 5 0., female 1,348 1,365 3 4 <0 5 5 <0., school levelc, middle school 1,148 1,136 4 4 0 5 6 0., high school 1,335 1,404 3 4 <0 5 5 <0., ethnicity/raced, white 540 499 4 4 0 5 6 <0., black 638 706 3 4 <0 5 5 0., hispanic 414 435 3 4 <0 5 5 0., asian 546 520 3 3 0 6 5 0., 98 92 3 4 0 5 5 0., mixed/other 293 288 4 4 0 5 5 0., socioeconomic statuse, low 936 973 3 3 0 5 5 0., low middle 560 556 3 4 <0 5 5 <0., middle 436 430 3 4 <0 5 5 0., high middle 335 320 4 4 0 5 5 0., high 199 193 4 5 0 5 6 0..

a The 1999 sample was weighted to allow for an examination of secular trends in meal patterns independent of demographic shifts in the population. For example, estimates of weekly breakfast frequency within the low socioeconomic status group in 1999 and 2010 are mutually controlled so that sex, school level, and ethnicity/race makeup are the same in the low socioeconomic status group in the 1999 sample as in the 2010 sample. b P values represent testing to examine weighted mean differences in meal frequency between 1999 and 2010. c Middle school represents students enrolled in 6th to 8th grades and high school represents students enrolled in 9th to 12th grades. d Adolescents could choose >1 ethnic/racial category; those responses indicating multiple categories were coded as “Mixed/Other.” Because there were few participants who identified themselves as Hawaiians or Pacific Islanders, these participants were also included in the “Mixed/Other” category. e. The prime determinant of socioeconomic status was the higher education level of either parent with adjustments made for student eligibility for free/reduced-price school meals, family public assistance receipt, and parent employment status. Reproduced from: “Secular Trends in Meal and Snack Patterns among Adolescents from 1999 to 2010,” by N. Larson, M. Story, M. E. Eisenberg, & D. Neumark-Sztainer, 2016, Journal of the Academy of Nutrition and Dietetics, 116 , 243. Reprinted with permission.

Descriptive Quantitative Designs 173

Descriptive statistics were used to show the percentage of children and adolescents, who did not eat lunch: 7% ± 1% (standard error) for 4- to 8-year-olds, 16% ± 2% for, 9- to 13-year-olds, and 17 ± 1% for 14- to 18-year-olds. linear regression was used to, show that missing lunch was associated with significantly lower intake of many micro-, nutrients for the day in all age groups., descriptive correlational design, correlation is a statistical procedure used to measure and describe the relationship or, association between two variables. the researcher may not know whether the variables, are related, or may suspect that one influences the other. in either case, no attempt is, made to manipulate an independent variable in correlational designs, so you cannot, conclude that the relationship is causal simply based on correlation., before a correlation coefficient can be calculated, you need to draw a scatterplot, with the quantitative data, as seen in figure 6. each dot in the scatterplot represents, one variable (x or y) from one person or observation. the values for the x variable are, placed on the x-axis (horizontal axis), and the values for the y variable are on the y-axis, (vertical axis)., the variable for the y-axis should be the outcome variable, which may also be, called the response or dependent variable. the variable for the x-axis should be the, predictor variable, which also may be called the explanatory or independent variable., for example, in a study on carbohydrate intake and dental caries, the researcher wants, to see if increasing carbohydrate intake (predictor or independent variable) increases the, number of cavities (outcome or dependent variable). so the number of cavities should, be on the y-axis and carbohydrate intake should be on the x-axis. this way, when you, look at the scatterplot, as the carbohydrate intake (predictor variable) increases along the, horizontal axis, you can see how it affects the number of cavities (outcome variable) on, the vertical axis., figure 6. 5 example of a scatterplot.

####### 240 280 320 360 400 440 480 560 600 640

####### 520

174 Chapter 6: Quantitative Research Designs: Experimental, Quasi-Experimental, and Descriptive

• Multiple Choice

Subject : Biology

• Discover more from: Biology Honors 999+   Documents Go to course
• More from: Biology Honors 999+   Documents Go to course

Quantitative Research Designs: Experimental, Quasi-Experimental, and Descriptive NOT for SALE OR DISTRIBUTION NOT for SALE OR DISTRIBUTION

Total Page： 16

File Type： pdf , Size： 1020Kb

• Abstract and Figures
• Public Full-text

• Statistical Inference: How Reliable Is a Survey? Math 203, Fall 2008: Statistical Inference: How reliable is a survey? Consider a survey with a single question, to which respondents are asked to give an answer of yes or no. Suppose you pick a random sample of n people, and you find that the proportion that answered yes isp ˆ. Question: How close isp ˆ to the actual proportion p of people in the whole population who would have answered yes? In order for there to be a reliable answer to this question, the sample size, n, must be big enough so that the sample distribution is close to a bell shaped curve (i.e., close to a normal distribution). But even if n is big enough that the distribution is close to a normal distribution, usually you need to make n even bigger in order to make sure your margin of error is reasonably small. Thus the first thing to do is to be sure n is big enough for the sample distribution to be close to normal. The industry standard for being close enough is for n to be big enough so that 1 − p 1 − p n > 9 and n > 9 p p both hold. When p is about 50%, n can be as small as 10, but when p gets close to 0 or close to 1, the sample size n needs to get bigger. If p is 1% or 99%, then n must be at least 892, for example. (Note also that n here depends on p but not on the size of the whole population.) See Figures 1 and 2 showing frequency histograms for the number of yes respondents if p = 1% when the sample size n is 10 versus 1000 (this data was obtained by running a computer simulation taking 10000 samples). [Show full text]
• On Becoming a Pragmatic Researcher: the Importance of Combining Quantitative and Qualitative Research Methodologies DOCUMENT RESUME ED 482 462 TM 035 389 AUTHOR Onwuegbuzie, Anthony J.; Leech, Nancy L. TITLE On Becoming a Pragmatic Researcher: The Importance of Combining Quantitative and Qualitative Research Methodologies. PUB DATE 2003-11-00 NOTE 25p.; Paper presented at the Annual Meeting of the Mid-South Educational Research Association (Biloxi, MS, November 5-7, 2003). PUB TYPE Reports Descriptive (141) Speeches/Meeting Papers (150) EDRS PRICE EDRS Price MF01/PCO2 Plus Postage. DESCRIPTORS *Pragmatics; *Qualitative Research; *Research Methodology; *Researchers ABSTRACT The last 100 years have witnessed a fervent debate in the United States about quantitative and qualitative research paradigms. Unfortunately, this has led to a great divide between quantitative and qualitative researchers, who often view themselves in competition with each other. Clearly, this polarization has promoted purists, i.e., researchers who restrict themselves exclusively to either quantitative or qualitative research methods. Mono-method research is the biggest threat to the advancement of the social sciences. As long as researchers stay polarized in research they cannot expect stakeholders who rely on their research findings to take their work seriously. The purpose of this paper is to explore how the debate between quantitative and qualitative is divisive, and thus counterproductive for advancing the social and behavioral science field. This paper advocates that all graduate students learn to use and appreciate both quantitative and qualitative research. In so doing, students will develop into what is termed "pragmatic researchers." (Contains 41 references.) (Author/SLD) Reproductions supplied by EDRS are the best that can be made from the original document. On Becoming a Pragmatic Researcher 1 Running head: ON BECOMING A PRAGMATIC RESEARCHER U.S. [Show full text]
• A Case-Crossover Study of Sleep and Childhood Injury A Case-Crossover Study of Sleep and Childhood Injury Francesca Valent, MD*; Silvio Brusaferro, MD*; and Fabio Barbone, MD, DrPH*‡ ABSTRACT. Objective. To evaluate the association n Italy (ϳ56 million inhabitants; 9 million chil- between sleep and wakefulness duration and childhood dren Ͻ15 years of age), ϳ500 children Ͻ15 years unintentional injury. of age are killed each year in unintentional inju- Study Design. Case-crossover study. I 1 ries. Homicides and suicides are much less frequent, Methods. Two hundred ninety-two injured children accounting for Ͻ10% of all fatal injuries among chil- who presented at the Children’s Emergency Center of dren.1 In Northern Italy, in 1995, injury mortality Udine, Italy, or their parents were interviewed after a rates (International Classification of Diseases, Ninth Re- structured questionnaire. Information was collected con- vision codes from 800 to 999) were 10 deaths/100 000 cerning sociodemographic variables, participant’s habits, and injury characteristics, including a brief description inhabitants under 1 year of age, 5/100 000 from 1 to of the accident dynamics. Sleep or wakefulness status of 4 years of age, and 6/100 000 from 5 to 14 years of the child was assessed retrospectively for each of the 48 age. Among infants, injury represented only 2% of hours before injury. For each child, we compared the 24 overall mortality, whereas in the 2 older age groups hours immediately before the injury (hours 1–24; case injuries accounted for 18% and 33% of all deaths, period) with hours 25 to 48 (control period). Nonparamet- respectively.2 Mortality, however, is just the tip of a ric tests were conducted to compare the difference of pyramid, where for each injury resulting in a death sleep duration between case and control periods. [Show full text]
• Those Missing Values in Questionnaires Those Missing Values in Questionnaires John R. Gerlach, Maxim Group, Plymouth Meeting, PA Cindy Garra, IMS HEALTH; Plymouth Meeting, PA Abstract alternative values are not bona fide SAS missing values; consequently, a SAS procedure, expectedly, will include Questionnaires are notorious for containing these alternative values, thereby producing bogus results. responses that later become missing values during processing and analyses. Contrary to a non-response that Typically, you can re-code the original data, so results in a typical, bona fide, missing value, questions that the missing values become ordinary missing values, might allow several alternative responses, necessarily thus allowing SAS to process only appropriate values. Of outside the usual range of appropriate responses. course, there's a loss of information since the original Whether a question represents continuous or categorical missing values did not constitute a non-response. Also, data, a good questionnaire offers meaningful alternatives, such pre-processing might include a myriad of IF THEN / such as: "Refused to Answer" and, of course, the ELSE statements, which can be very tedious and time- quintessential "Don't Know." Traditionally, these consuming to write, and difficult to maintain. Thus, given alternative responses have numeric values such as 97, a questionnaire that has over a hundred variables with 998, or 9999 and, therefore, pose problems when trying varying levels of missing values, the task of re-coding to distinguish them from normal responses, especially these variables becomes very time consuming at best. when multiple missing values exist. This paper discusses Even worse, the permanent re-coding of alternative missing values in SAS and techniques that facilitate the responses to ordinary missing numeric values in SAS process of dealing with multi-valued, meaningful missing precludes categorical analysis that requires the inclusion values often found in questionnaires. [Show full text]
• 10 Questions Opinion Polls Questions you might have on 10opinion polls 1. What is an opinion poll? An opinion poll is a survey carried out to measure views on a certain topic within a specific group of people. For example, the topic may relate to who Kenyans support in the presidential race, in which case, the group of people interviewed will be registered voters. 2. How are interviewees for an opinion poll selected? The group of people interviewed for an opinion poll is called a sample. As the name suggests, a sample is a group of people that represents the total population whose opinion is being surveyed. In a scientific opinion poll, everyone has an equal chance of being interviewed. 3. So how come I have never been interviewed for an opinion poll? You have the same chance of being polled as anyone else living in Kenya. However, chances of this are very small and are estimated at about 1 in 14,000. This is because there are approximately 14 million registered voters in Kenya and, for practical and cost reasons, usually only between 1,000 and 2,000 people are interviewed for each survey carried out. 4. How can such a small group be representative of the entire population? In order to ensure that the sample/survey group is representative of the population, the surveyors must ensure that the group reflects the characteristics of the whole. For instance, to get a general idea of who might win the Kenyan presidential election, only the views of registered voters in Kenya will be surveyed as these are the people who will be able to influence the election. [Show full text]
• SAMPLING DESIGN & WEIGHTING in the Original Appendix A 2096 APPENDIX A: SAMPLING DESIGN & WEIGHTING In the original National Science Foundation grant, support was given for a modified probability sample. Samples for the 1972 through 1974 surveys followed this design. This modified probability design, described below, introduces the quota element at the block level. The NSF renewal grant, awarded for the 1975-1977 surveys, provided funds for a full probability sample design, a design which is acknowledged to be superior. Thus, having the wherewithal to shift to a full probability sample with predesignated respondents, the 1975 and 1976 studies were conducted with a transitional sample design, viz., one-half full probability and one-half block quota. The sample was divided into two parts for several reasons: 1) to provide data for possibly interesting methodological comparisons; and 2) on the chance that there are some differences over time, that it would be possible to assign these differences to either shifts in sample designs, or changes in response patterns. For example, if the percentage of respondents who indicated that they were "very happy" increased by 10 percent between 1974 and 1976, it would be possible to determine whether it was due to changes in sample design, or an actual increase in happiness. There is considerable controversy and ambiguity about the merits of these two samples. Text book tests of significance assume full rather than modified probability samples, and simple random rather than clustered random samples. In general, the question of what to do with a mixture of samples is no easier solved than the question of what to do with the "pure" types. [Show full text]
• The Very Process of Taking a Pill May Create Expectations in a Patient Which May Affect Reactions to the Pill Ch 1 Pause for Thought Questions p. 36 1- Why is it important to use placebos and a double-blind approach in some studies? The very process of taking a pill may create expectations in a patient which may affect reactions to the pill. To help eliminate the role of patient expectation during testing, a placebo (or fake medicine) is used on subjects in a control group. A researcher’s expectations may also affect results. Therefore, double-blind studies are often used. In such studies, neither the researcher nor the participants know which group is which. 2- Assume that researchers find that people’s memories are sharpest right after they’ve eaten lunch. What hidden variables may have affected these results? Other variables that may play a role (hidden variables): Time of Day Type of food eaten Meaning of “lunch” - May mean “time away from work” - Subjective State of feeling free (not food) helps memory. Socializing at lunch may make a person more alert and then able to take-in more info and retain it longer. 3- How might you use the scientific method to study factors that affect obedience? Devise a simple study, and identify the following: hypothesis, subjects, independent variable, dependent variable, experimental group, and control group. Hypothesis: If employee fears/believes they will be punished if they don’t follow directives, then they will be obedient to directives of bosses/people in “higher” positions than themselves in the workplace. Subjects: Employees Independent Variable: Directive given with harsh consequences for not following through. Dependent Variable: Obedience (Level of) Experimental Group: Group with “harsh” bosses (very by the book; give extreme consequences) present during their work day. [Show full text]
• Introduction 1 Introduction he field of public health has never been as widely known or popular as T in recent years. On a global scale, the spread of HIV/AIDS beginning in the 1980s gave public health enormous impetus and visibility. Much like infectious diseases from earlier eras, HIV/AIDS was deeply enmeshed in environmental and behavioral contexts. If left unaddressed, the disease promised to engulf large portions of the world’s population. Yet today’s most enduring and pervasive public health problems are far more mundane, e.g., poor sanitation and water quality, malnutrition, and the everyday violence of grinding poverty. The 20th-century reign of the germ theory of disease etiology, with its emphasis on curing over preven- tion and laboratories over communities, has been tempered by these real- ities and by the vast increase in chronic diseases such as hypertension, diabetes, and cancer. Similarly, the dominance of quantification, in which ever-more sophisticated measures and statistics are expected to capture the full range of human experience, has given way to a more nuanced and thoughtful matching of methods with the problem at hand as well as with the people and places experiencing it (Baum, 1996; Rapkin & Trickett, 2005). Enter Qualitative Methods A colleague once astutely remarked that virtually anyone can read and appreciate qualitative research—its narrative reporting style makes it appear easy to carry out. By comparison, a quantitative study relies on complicated 1 2 QUALITATIVE AND MIXED METHODS IN PUBLIC HEALTH statistical analyses that require prior knowledge to decode their meaning. Yet the appealing end product of a qualitative study represents the culmi- nation of intense involvement and intellectual labor. [Show full text]
• Chapter 5 Contrasts for One-Way ANOVA Page 1. What Is a Contrast? Chapter 5 Contrasts for one-way ANOVA Page 1. What is a contrast? 5-2 2. Types of contrasts 5-5 3. Significance tests of a single contrast 5-10 4. Brand name contrasts 5-22 5. Relationships between the omnibus F and contrasts 5-24 6. Robust tests for a single contrast 5-29 7. Effect sizes for a single contrast 5-32 8. An example 5-34 Advanced topics in contrast analysis 9. Trend analysis 5-39 10. Simultaneous significance tests on multiple contrasts 5-52 11. Contrasts with unequal cell size 5-62 12. A final example 5-70 5-1 © 2006 A. Karpinski Contrasts for one-way ANOVA 1. What is a contrast? • A focused test of means • A weighted sum of means • Contrasts allow you to test your research hypothesis (as opposed to the statistical hypothesis) • Example: You want to investigate if a college education improves SAT scores. You obtain five groups with n = 25 in each group: o High School Seniors o College Seniors • Mathematics Majors • Chemistry Majors • English Majors • History Majors o All participants take the SAT and scores are recorded o The omnibus F-test examines the following hypotheses: H 0 : µ1 = µ 2 = µ3 = µ 4 = µ5 H1 : Not all µi 's are equal o But you want to know: • Do college seniors score differently than high school seniors? • Do natural science majors score differently than humanities majors? • Do math majors score differently than chemistry majors? • Do English majors score differently than history majors? HS College Students Students Math Chemistry English History µ 1 µ2 µ3 µ4 µ5 5-2 © 2006 A. [Show full text]
• Uses for Census Data in Market Research Uses for Census Data in Market Research MRS Presentation 4 July 2011 Andrew Zelin, Director, Sampling & RMC, Ipsos MORI Why Census Data is so critical to us Use of Census Data underlies almost all of our quantitative work activities and projects Needed to ensure that: – Samples are balanced and representative – ie Sampling; – Results derived from samples reflect that of target population – ie Weighting; – Understanding how demographic and geographic factors affect the findings from surveys – ie Analysis. How would we do our jobs without the use of Census Data? Our work withoutCensusData Our work Regional Assembly What I will cover Introduction, overview and why Census Data is so critical; Census data for sampling; Census data for survey weighting; Census data for statistical analysis; Use of SARs; The future (eg use of “hypercubes”); Our work without Census Data / use of alternatives; Conclusions Census Data for Sampling For Random (pre-selected) Samples: – Need to define stratum sample sizes and sampling fractions; – Need to define cluster size, esp. if PPS – (Need to define stratum / cluster) membership; For any type of Sample: – To balance samples across demographics that cannot be included as quota controls or stratification variables – To determine number of sample points by geography – To create accurate booster samples (eg young people / ethnic groups) Census Data for Sampling For Quota (non-probability) Samples: – To set appropriate quotas (based on demographics); Data are used here at very localised level – ie Census Outputs Census Data for Sampling Census Data for Survey Weighting To ensure that the sample profile balances the population profile Census Data are needed to tell you what the population profile is ….and hence provide accurate weights ….and hence provide an accurate measure of the design effect and the true level of statistical reliability / confidence intervals on your survey measures But also pre-weighting, for interviewer field dept. [Show full text]
• Survey Experiments IU Workshop in Methods – 2019 Survey Experiments Testing Causality in Diverse Samples Trenton D. Mize Department of Sociology & Advanced Methodologies (AMAP) Purdue University Survey Experiments Page 1 Survey Experiments Page 2 Contents INTRODUCTION ............................................................................................................................................................................ 8 Overview .............................................................................................................................................................................. 8 What is a survey experiment? .................................................................................................................................... 9 What is an experiment?.............................................................................................................................................. 10 Independent and dependent variables ................................................................................................................. 11 Experimental Conditions ............................................................................................................................................. 12 WHY CONDUCT A SURVEY EXPERIMENT? ........................................................................................................................... 13 Internal, external, and construct validity .......................................................................................................... [Show full text]
• CORRELATION COEFFICIENTS Ice Cream and Crimedistribute Difficulty Scale ☺ ☺ (Moderately Hard)Or 5 COMPUTING CORRELATION COEFFICIENTS Ice Cream and Crimedistribute Difficulty Scale ☺ ☺ (moderately hard)or WHAT YOU WILLpost, LEARN IN THIS CHAPTER • Understanding what correlations are and how they work • Computing a simple correlation coefficient • Interpretingcopy, the value of the correlation coefficient • Understanding what other types of correlations exist and when they notshould be used Do WHAT ARE CORRELATIONS ALL ABOUT? Measures of central tendency and measures of variability are not the only descrip- tive statistics that we are interested in using to get a picture of what a set of scores 76 Copyright ©2020 by SAGE Publications, Inc. This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher. Chapter 5 ■ Computing Correlation Coefficients 77 looks like. You have already learned that knowing the values of the one most repre- sentative score (central tendency) and a measure of spread or dispersion (variability) is critical for describing the characteristics of a distribution. However, sometimes we are as interested in the relationship between variables—or, to be more precise, how the value of one variable changes when the value of another variable changes. The way we express this interest is through the computation of a simple correlation coefficient. For example, what’s the relationship between age and strength? Income and years of education? Memory skills and amount of drug use? Your political attitudes and the attitudes of your parents? A correlation coefficient is a numerical index that reflects the relationship or asso- ciation between two variables. The value of this descriptive statistic ranges between −1.00 and +1.00. [Show full text]

• Request new password
• Create a new account

Doing Research in the Real World

Student resources, chapter 6: research design: quantitative methods.

• Checklist for Using Theory