Chapter 1114 The Nuts and Bolts of Experiments

Chapter 11/14 The Nuts and Bolts of Experiments Statistics (Means) Dawn M. Mc. Bride The Process of Research in Psychology Second Edition

Steps in the Research Process 2 Dawn M. Mc. Bride - The Process of

Experiments: Essential Elements l Manipulation of an independent variable l Treatment groups formed by levels or types of independent variable l Random assignment of participants to groups l Control for confounding variables (3 rd variable P. ) § 3 Characteristics of settings or participants that could explain the observed results (dependent variable) Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Experiments l Independent Variable§ variable that is manipulated by the researcher § 4 researcher controls the administration of the levels of the variable to the participants. Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Independent variables can be manipulated in two ways: l l Between-subjects: § Each subject receives only one level of the IV § Subjects are randomly assigned to levels Within-subjects: § 5 Each subject receives all levels of the IV Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Between-Subjects Matched Designs: 6 l Concern: § There may be individual differences across groups that may affect the outcome of interest (dependent variable) l Random assignment may not be sufficient to control for these differences (particularly if the sample is small) Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Between-Subjects Matched Designs: l Participants are matched on a characteristic that may contribute to group differences. § 7 Gender, marital status, education, ethnicity, level of depression, attitudes, beliefs l May need to use a pre-test (e. g. depression) l Each participant within a matched set is randomly assigned to treatment groups. Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Random Assignment Do students who take a practice exam do better than students who do not take a prac. exam? 8

Matched Random Assignment Will year in school affect study results? 9

Within-subjects Designs 10 l Each participant receives all treatments l Each participant is its own control l It is possible that order of the levels can affect the dependent variable l Counterbalancing- manipulate order of the levels Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Within-subjects Designs: Counter-balancing 11 .

Factorial Designs 12 l Include more than one independent variable l Allow to test for the effects of multiple independent variables in one experiment l And examine the effects of the interactions of those independent variables on the dependent variable. Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Factorial Design: Is the effect different for boys than girls? (1 True IV; 1 Quasi-I. IV) 13 .

Factorial Designs with 2 IVs: l Two main effects, one for each IV § § l Gender: Boys Girls (moderator) Interaction effect § 14 Type of video watched: Reward, Punish, No Consequence Results when taking into account both IVs together Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Factorial Designs: Interpreting Interactions 15 l Graphs are useful in determining the type of interaction that occurred l Follow-up statistical tests to examine interactions are called simple effects tests. Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Example Factorial Designs: 2 true IVs Does method of study words: IV#A – Study Mode Auditory vs. Visual And method of recall words: IV#B- Test Mode Auditory vs. Visual Make a difference in DV-Word recall test ? 16 Text Pages 230 -233

Conclusion from main effects? Means for Main Effects: A 1 A 2 IV#A – Study Mode Auditory vs. Visual 62. 5 67. 5 IV#B- Test Mode Mean 17 B 1 B 2 Auditory vs. Visual 67. 5 62. 5

Factorial Designs- Between Groups 18 Dawn M. Mc. Bride - The Process of Research

Interactions Possibility 1: Best when study and visual method match 19 Dawn M. Mc.

Interaction Possibility 2: Type of test makes difference for auditory study only 20 Dawn

Interaction Possibility #3 : Auditory test best regardless of study type 21 Dawn M.

Interaction- C. Chips & Vanilla Flavor- Ratings range 1 -7 22 Dawn M. Mc.

Interaction- C. Chips & Vanilla Flavor- Ratings range 1 -7 23 Dawn M. Mc.

Lab 27 # 1 ---- 2 x 2 100 90 80 70 60 50 40 30 20 10 0 Male S Female S Bystander 0 24 Bystander 10 Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Lab 27 # 2 --- 3 x 2 25 Dawn M. Mc. Bride -

Lab 28 a. 600 500 400 Group 300 Indiviudal 200 100 0 1 Week

Lab 28 b. 800 700 600 500 Group 400 Indiviudal 300 200 100 0

Lab 28 c. 450 400 350 300 250 Group 200 Indiviudal 150 100 50

Lab 28 d. 450 400 350 300 250 Group 200 Indiviudal 150 100 50

Using Statistics l 30 Hypothesis Testing § Hypothesis tests are designed to look for evidence against the null hypothesis § If calculated statistic falls in the critical region, of possible values if null is true, then we have enough evidence to: Ø reject the null hypothesis and Ø accept the scientific/alternative hypothesis that there is a difference or a relationship. Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Using Statistics l Comparing l Testing 31 Means Relationships Dawn M. Mc. Bride -

32 Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd

Significance Testing: Means 33

Variability and the T-Test 34 Dawn M. Mc. Bride - The Process of Research

Comparing Means 35 l A significant difference between groups: reject the null hypothesis. l Sample mean differences can be caused by the effect, by sampling error, or both. l A larger statistical value (T statistic) indicates a larger treatment effect l Larger statistical values lead to more power to detect effects with the test Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

T- Test: One IV and One DV l Independent l One Sample (vs. a

Comparing Means Independent Samples T-test 37 l Independent samples t-test - is used when two samples of different individuals are compared. (Between Groups) l Each sample represents the population for that group or condition. (between groups) § Depression scores among men and women § Depression scores among treated and control group Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Formula for T-Test (Mean, SD and N – sample size) 38 Dawn M. Mc.

SE: Standard error of the difference SD 2 = Variance 39 Dawn M. Mc.

Comparing Means: One-sample t-test l 40 Used to compare one samples' score to a known “population” mean (from standardized scores; or to chance) § Compare 8 th grade students mean score in the algebra standardized test to the average score in the standardized algebra test § Participants in the subliminal treatment group will recognize the images more often than 50% of the time (book example p. 287) Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Comparing Means Paired/Related Samples T-test l Scores from two related samples are compared § Relationship satisfaction among spouses or l Two sets of scores from the same individual (within groups) § l 41 Depression scores before and after treatment In these cases, variability across samples is likely to be lower than if samples are independent.

T- Test: One IV and One DV l Used to compare only two means (1 IV & 1 DV) § § § l Independent Samples One Sample (vs. a known value) Paired Samples Can be used in experimental designs with § § one IV that only has two levels: Ø 42 e. g. Treatment vs. Control Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Factorial Design: Is the effect different for boys than girls? (1 True IV; 1 Quasi-I. IV) 43 .

How many T-tests (. 05 x 3 =. 15 ) l For main effect treatment (mixing boys and girls) Reward 44 Punish No Consequence Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Comparing Means: ANOVA l An IV with more than two levels § Type of model: Reward, Punish, No Conseq. and l Factorial Anova: more than one IV § § l Type of model: Reward, Punish, No C Gender: male vs. female The comparison of means for each level of the IV (true or quasiindependent) variable is called the main 45 effect. Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

Factorial Anova: more than one IV 2 X 2 Means for Main Effects: A 1 A 2 IV#A – Study Mode Auditory vs. Visual 62. 5 67. 5 IV#B- Test Mode Mean 46 B 1 B 2 Auditory vs. Visual 67. 5 62. 5

ANOVA Design # Ivs & # levels 2 x 2 IV 1 IV 2 Treatment: Gender: CBT vs. control male vs. female Treatment: Gender: CBT vs Interpersonal vs. Control 3 x 2 IV 1 IV 2 47 male vs. female Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

ANOVA Design 2 x 2 x 2 IV 1 IV 2 IV 3 Teaching method: Auditory vs Visual Recall Method: Auditory vs Visual Gender: male vs. female 2 x 2 x 2 x 3 IV 1 IV 2 IV 3 IV 4 48 Teaching method: Recall Method: Gender: Grade Auditory vs Visual male vs, female 6, 7, 8

Comparing Means: ANOVA l Interaction Effect § 49 The effect of one independent variable (treatment) at each level of the other variable (gender). moderator Ø Was the treatment more effective than the control group for males? Ø Was the treatment more effective than the control group for females ? Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.

ANOVA vs. Test 2 x 2 l l ANOVA more stringent test Alpha of less than. 05 to detect the presence of § § l Simple Effects § 50 Main effects Interaction Effect If the ANOVA detects an effect, then T-tests are conducted to see where are the differences Dawn M. Mc. Bride - The Process of Research in Psychology, 2 nd Edition © 2013 SAGE Publications, Inc.