Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
46 Cards in this Set
- Front
- Back
Debriefing
|
informing the participant of the true nature of the study, and should occur immediately after data collection if possible. • Sometimes the investigator will provide the participant with a brief write-up explaining the hypothesis, the significance of the study, and what he hopes to find. • This also serves to help the participant return to their previous state of mind.
|
|
Understand “Risk” and minimal risk
|
Risk – participants are considered at risk if their privacy is compromised or they are placed in a situation that more physical or emotional risk than they would experience in their routine daily lives or in routine physical or psychological examinations.. • Minimal Risk – the participants are placed under no more physical or psychological stress than that encountered in their daily lives or in routine physical or psychological examinations.
|
|
How do you write in APA format: t obt is 2.41 the df is 16 and p is less than .01?
|
T(16) = 2.60, p < 0.01
|
|
between participants design
|
no same person participates in the control and experimental groups.
|
|
What do you have to control in a between participants exp.
|
"1.) Who is in the study. Remember we want a sample that is representative of the population. 2.) Which condition (or group) the participants are assigned to. We should randomly assign the participants to the different conditions. 3.) What happens in the experiment, so the only difference between the 2 conditions is the level of the independent variable"
|
|
what is a disadvantage to the pretest/post test design
|
this between subjects design may increase demand characteristics and experimenter effects.
|
|
threats to internal validity
|
Nonequivalent control group
History Maturation Testing Regression to the mean Instrumentation Mortality and attrition Diffusion of treatment Experimenter and participant effects Floor and ceiling effects |
|
what do blinded experiments control for
|
participant and experimenter effects
|
|
Matched participants design
|
Type of Correlated-Group Designs: • Matched-participants experimental design• This is a cross between the between- and within-participants designs. • In this design not all participants partake in each condition. Like the between-participants design, half the participants are assigned to the experimental condition and the other half to the control condition. • For each participant in one condition, there is another participant in the other condition that matches him/her on the relevant variables. Advantages: • Eliminates carryover effects. • Minimizes testing effects and demand characteristics • The groups are more equivalent than those in the between-participant design and almost as equivalent as those in the within-participant design • Because the participants have been matched on the variable(s) of interest, the same statistics used for the within-participants design can be used on the matchedparticipants design. Disadvantages: • More participants are needed than in the withinparticipants design. • If one participant drops out, the entire pair is lost. • Matching participants can be VERY difficult.
|
|
Problems with external validity
|
Generalizations to populations - College Sophmore problem. Generalization from Laboratory Settings
|
|
Correlated-Group Designs
|
Within-participants experimental design – the same participants are used in each condition.
Matched-participants design – the participants are matched between conditions or variables that the experimenter believes are (is) relevant to the study. |
|
What is another name for Within design
|
repeated measures
|
|
Advantages to Within design
|
Randomly recruit participants but don’t need random assignment because all participants will partake in both conditions.
Typically requires fewer participants than between-participants design. Requires less time to conduct Increases statistical power. When the same individuals participate in multiple conditions, individual differences between conditions are minimized. This reduces variability |
|
Disadvantages to Within design and ways to overcome them
|
Internal validity is a concern.
Order effects (another type of testing effect) are of concern because all participants are measured at least twice, so the order in which the conditions are presented may have an effect on the dependent variable. Practice and Fatigue effects Counterbalancing , systematically varying the order of the conditions for participants in a within-participants design, can control for these potential confounding effects. |
|
Matched-participants experimental design benefits
|
Eliminates carryover effects.
Minimizes testing effects and demand characteristics The groups are more equivalent than those in the between-participant design and almost as equivalent as those in the within-participant design Because the participants have been matched on the variable(s) of interest, the same statistics used for the within-participants design can be used on the matched-participants |
|
Matched-participants experimental design disadvantages
|
More participants are needed than in the within-participants design.
If one participant drops out, the entire pair is lost. Matching participants can be VERY difficult. |
|
Review slide 35 on review page
|
go ahead and do it you know you want to
|
|
3 types of developmental designs
|
1.) Cross-sectional design
2.) Longitudinal design 3.) Sequential design |
|
Cross-sectional design
|
Participants of different ages are studied at the same time.
Advantage: Fast – a wide variety of ages can be studied in a short period of time. Disadvantage: Cohort effects – (cohort = a group of individuals born at the same time) – With the cross-sectional design the researcher wants to conclude that any observed difference are due to age, but they could also be due to cohort effects, the differences that may arise because the people were raised at different times and in different generations (or cohorts). |
|
Longitudinal design
|
The same participants are studied repeatedly over time (follow the participants over time).
Advantage: No cohort effects Disadvantage: Expensive Time consuming Attrition |
|
Sequential design
|
Combination of the cross-sectional and longitudinal designs.
Researcher begins with participants of differing ages (cross-sectional design) and tests or measures them. And then follows them over a period of time (longitudinal design). Advantage: Allows researchers to examine cohort effects. Disadvantage: More expensive and time consuming than the previous two designs. Attrition |
|
Hypothesis testing
|
process of determining whether a given hypothesis is supported by the experimental results.
|
|
Inferential statistics
|
procedures used for drawing conclusions about a population based on the data collected from the sample. attempts to deduce the properties of a large population based on the results from a small sample of that population.
- We use probability values (p-values) to determine statistical significance |
|
Descriptive Statistics
|
mathematical summaries of results. There are two broad categories of descriptive statistics:
-Measurements of the central score (i.e., mean, median, and mode). -Measurements of variation or dispersion (i.e., range, average deviation and standard deviation). |
|
Alternative hypothesis
|
hypothesis that the researcher wants to support (i.e., the hypothesis of interest). That is, the hypothesis that predicts that there is a difference between the groups of interest
|
|
Type I error
|
null hypothesis is rejected when it is true. (false positive)
|
|
Type II error
|
null hypothesis was accepted when it was false. (false negative)
|
|
At p≤0.05, we are stating that there is 95% chance that our results are due to the experimental manipulation and 5% chance that they are due to error. What type of error is that 5% (i.e., how is that error classified)?
|
Type I error
|
|
Single-group design and test you can use
|
Can use the z test. This design lacks a comparison (or control group). We can compare the performance of the sample with that of the population, assuming the population data are available.
|
|
Parametric test
|
involves making assumptions about estimates of the population characteristics, or parameters. These assumptions involve knowing the mean and the standard deviation of the population and that the population distribution is normal.
examples t and z tests |
|
Nonparametric test
|
test that does not involve the use of any population parameters (e.g., the mean or standard deviation) and the underlying distribution does not need to be normal.
examples chi squared and Wilcoxon tests |
|
z test formula
|
z = (Xbar - mu)/SIGMAxbar (meaning standard error)
|
|
diff between one and 2 tailed z tests
|
one tailed test alpha is 1.645 and 2 tailed alpha is 1.96
|
|
Ways of increasing stat. power with z test
|
use 1 tail, or increase N
|
|
assumptions for z test
|
1.) μ (mean of sampling distribution) and σ (standard deviation) are known.
2.) Appropriate for interval or ratio data. 3.) The distribution of the random samples is normal (bell-shaped). 4.) The sample size is not too small. Why not small samples? Because often times they fail to form a normal distribution (not enough data points). |
|
CI
|
CI = X ± z (σx) where,
X = the sample mean σx = the standard error of the mean z = the z-score representing the desired CI |
|
t test definition
|
The t test is a parametric inferential statistical test of the null hypothesis for a single sample where the population variance is NOT known. T distributions are symmetrical and bell-shaped but they do not fit the standard normal distribution.
|
|
formula for t test
|
(Xbar - mu)/ standard error using s (instead of sigma xbar)
|
|
what is s
|
estimated standard dev of pop this is the formula you know using N-1
|
|
independent groups t test formula
|
Not matched design
(Xbar1-2)/(SE of Xbar1-2) SE of Xbar1-2 = SQRT ((s1 squared / n1) + (s2 squared / s2)) df = (n1 – 1) + (n2 - 1) |
|
effect size
|
proportion of variance in the dependent variable that is accounted for by the manipulation of the independent variable.
larger the effect size, the more consistent is the influence of the independent variable on the dependent variable. |
|
how do you measure effect size
|
use cohen's d
same formula as t statistic only using 2 in the denom of the se formula instead of n1 and n2 Small effect size at least 0.02; Medium effect size is at least 0.05; Large effects size is at least 0.08. |
|
assumptions of independent t test
|
1.) Data are interval or ratio.
2.) The underlying distributions are bell-shaped. 3.) The observations are independent. 4.) Homogeneity of variance – if we could compute the true variance of the population represented by each sample, the variances in each population would be the same.” |
|
Standard error of the difference scores
|
the standard deviation of the sampling distribution of the mean differences between dependent samples in a two groups design
|
|
df for t test for correlated groups
|
n-1
|
|
Assumptions of the Correlated-Groups t test:
|
1.) Data are interval or ratio.
2.) The underlying distributions are bell-shaped. 3.) The observations are NOT independent. 4.) Homogeneity of variance – “if we could compute the true variance of the population represented by each sample, the variances in each population would be the same.” |