Psyc 450 Lab Homework #3

Psyc 450 Lab Homework #3

Psyc 450 Lab Homework #3

Walk Through use the “Juvenile” data set

Research hypotheses: -- DV is Extreme verbal abuse during the 4th month of treatment (pa4)

RH1: Males will have lower verbal abuse means than females

RH2: Those with prior criminal offenses will have higher verbal abuse means than those with no prior offenses

RH3: There will be an interaction, with a prior offense effect for females but not for males.

Thinking about the hypotheses…

  • Label the boxes to depict this research design.
  • Tell the two sets of simple effects involved in this design
  • Tell which set of simple effect will directly test the interaction research hypothesis
  • Put in the <, > or = to depict the simple effects that describe the interaction
  • Put in the <, > or = to depict the main effect of Gender
  • Does the interaction RH: suggest this main effect will be descriptive or misleading?
  • Put in the <, > or = to depict the main effect of Prior Criminal Offenses
  • Does the interaction RH: suggest this main effect will be descriptive or misleading?

Doing the analysis …

Obtain the analysis and complete the table on the right.

This analysis is based on a very small sample (recall that n=41 is about right to have 80% power for r=.30). So, this is a great opportunity to use effect size estimates to temper our usual reliance upon p-values!

While these p-values don’t actually quite meet the p < .05 standard, they are close, the sample size is small, the effects are fairly large & have the hypothesized pattern. Under these circumstances it is not uncommon for researchers to interpret the effects carefully and then emphasize the need for replication with a larger sample size.

Interaction F= ______df = ____, ______MSe = ______p = ______, r = ______

Computing the LSDmmd k = ______n = ______dferror = ______MSerror = ______LSDmmd = ______

Label & compute the effect size for each set of simple effects

Simple effect of ______for each level of ______

Simple effect of ______for ______r = ______

Simple effect of ______for ______r = ______

Simple effect of ______for each level of ______

Simple effect of ______for ______r = ______

Simple effect of ______for ______r = ______

Considering the simple effect sizes, are you “content” with the LSD results ? Why or why not?

Does the pattern of cell means match the RH? Completely Partially Not at all Briefly explain your answer.

Main effect of Gender F= ______df = ____, ______MSe = ______p = ______, r = ______

Considering the effect size, are you “content” with the main effect significance test? Why or why not?

Does the pattern of marginal means match the RH? Completely Partially Not at all Briefly explain your answer.

Now consider the simple effects of gender.

Is this main effect RH: supported? Completely Partially Not at all Briefly explain your answer.

Main effect of Prior Criminal Offense F= ______df = ____, ____ MSe = ______p = ______, r = ______

Considering the effect size, are you “content” with the main effect significance test? Why or why not?

Does the pattern of marginal means match the RH? Completely Partially Not at all Briefly explain your answer.

Now consider the simple effects of prior criminal offense.

Is this main effect RH: supported? Completely Partially Not at all Briefly explain your answer.

Write up the three effects, attending to the RH:s & your concerns about basing the results only upon the p-values.
Your Turn

  1. Using the clinical.sav dataset…

Research hypotheses:

RH1: Males will report lesser amounts of social support than women.

RH2: Married respondents will report greater social support than single respondents.

RH3: The difference in reported social support between men and women will be larger for married than for single repsondents.

Thinking about the hypotheses…

  • Label the boxes to depict this research design.
  • Tell the two sets of simple effects involved in this design
  • Tell which set of simple effect will directly test the interaction research hypothesis
  • Put in the <, > or = to depict the simple effects that describe the interaction
  • Put in the <, > or = to depict the main effect of Gender
  • Does the interaction RH: suggest this main effect will be descriptive or misleading?
  • Put in the <, > or = to depict the main effect of Marital Status
  • Does the interaction RH: suggest this main effect will be descriptive or misleading?

Getting ready to do the analysis…

The Analysis…

Complete the analysis for each of these three DVs

  • soss -- significant other social support
  • frss -- friend social support
  • fass -- family social support

Write up your results (include the effect size estimate for each main, interaction and simple effect you mention in the write-up) including nice bar graphs for each of the DVs – focus on the RH: support, non-support and consistency of results across the three DVs.

  1. Using the teenss.sav data set …

RH1: Male teenagers report engaging in more “mean” social activities than female teenagers.

RH2: “Popular” and “rejected” teenagers report engaging in the same amount of “mean” social activities.

Rh3: Rejected females report engaging in more “mean” social activities than popular females, whereas popular males report engaging in more “mean” activities than rejected males.

Thinking about the hypotheses…

  • Label the boxes to depict this research design.
  • Tell the two sets of simple effects involved in this design
  • Tell which set of simple effect will directly test the interaction research hypothesis
  • Put in the <, > or = to depict the simple effects that describe the interaction
  • Put in the <, > or = to depict the main effect of Gender
  • Does the interaction RH: suggest this main effect will be descriptive or misleading?
  • Put in the <, > or = to depict the main effect of Sociometric Status
  • Does the interaction RH: suggest this main effect will be descriptive or misleading?

Getting ready to do the analysis…

The Analysis…

Complete the analysis for each of these three DVs

  • i5 -- make jokes when others are clumsy
  • i12 – laugh at others mistakes
  • i37 – call bad names

Write up your results (include the effect size estimate for each main, interaction and simple effect you mention in the write-up) including nice bar graphs for each of the DVs – focus on the RH: support, non-support and consistency of results across the three DVs.