Controls to Reduce Threats to Validity and intro to statistics: The normal distribution

Question 1

what is controlling?

Answer 1

a procedure aimed to reduce potential threats to the validity. There are four types of control procedures.

Question 2

What are the three parts of”general control procedure”?

Answer 2

1. Preparation of the setting:structuring the research to eliminate confounding variable, simplify the situation and increase the control over the independent variable. –> Increasement in internal and external validity.
2. Response Measurement: selecting and preparing measuring instruments based on reliability and validity (increases statistical and construct validity)
3. Replication

Question 3

What are the three types of replication?

Answer 3

1. Exact replication
2. systematic replication: variation of the procedures in a systematic way, small changes based on theory
3. conceptual replication: replicating the original study, while using differently operationalized constructs.

Question 4

How can we control both, subject and experimenter effect?

Answer 4

1. Double blind procedures

2. automation

Question 5

How can we control subject effects?

Answer 5

Single Blind procedures

Question 6

What is the advantage of “objective measuring scales”?

Answer 6

that we dont have differences in results across different experimenters using the instrument

Question 7

multiple researchers show e.g. 85% agreement in their results, what can beused to measure this inter-rater relationship

Answer 7

“kappa” (cohen’s kappa coeffficient) It is generally thought to be a more robust measure than simple percent agreement calculation, since κ takes into account the possibility of the agreement occurring by chance.

Question 8

Deception

Answer 8

intentionally deceiving. can be done by using a placebo

Question 9

what is the balanced placebo design?

Answer 9

it’s a 2x2 factorial design, with (1) receiving treatment/not receiving treatment and (2) being told they are receiving placebo/not being told they are receiving a treatment…can reduce expectancy effect, but requires a debriefing after the study.

Question 10

From what population can the sample be selected?

Answer 10

From the available population, which is part of the target population (populationof interest to the stdy) , which is part of the general population.

Question 11

What are the ways to select a reprentative sample?

Answer 11

1. Random sampling 2. stratified random sampling: seprate sampling from different sub-populations. 3. Ad hoc sampling: creating a sample from the available population

Question 12

Another way to control? :)

Answer 12

multiple observers

Question 13

Ad hoc samples

Answer 13

Random and stratified
random samplings from a target population are important in some research, such as in political
surveys, but most psychological research does not employ random selection. Instead, psychologists
usually draw participants from an accessible population, such as introductory psychology students
or children from local schools. This type of sample, called an ad hoc sample, is used in most
psychological research. Furthermore, ethical considerations require that participants be volunteers.

Question 14

free random assignment

Answer 14

participants are randomly asigned to the research conditions, also contributing to the Initial Equivalence of the groups.

Question 15

Blocked random sampling

Answer 15

When we randomize within blocks, we use random assignment

to allocate blocks of participants to each of the available groups.

Question 16

Matched random assignment

Answer 16

This procedure involves matching
participants on a relevant variable, such as age or IQ, and then randomly assigning the matched
participants to groups, with one matched participant per group. (especially used in small samples)

Question 17

What can be a disadvantage of keeping a variable constant by matching? (control a variable)

Answer 17

One disadvantage to matching by holding the variable constant is that it reduces external
validity because it reduces the ability to generalize the results of the study to the larger population.
For example, if you used only adult participants, you would be unable to generalize with confidence
to adolescents.

Question 18

In random sampling: What variables can be controlled and which can not (or only hardly)

Answer 18

you can control some subject effects but usually not those that are environmental and setting related variables, such as time of day, room temperature, or background
noise. For these variables, the general control procedures discussed earlier are used.

Question 19

Control through experimental design:

Answer 19

Routinely including these two basic control procedures—(1) inclusion of appropriate
control groups or conditions, and (2) random assignment of participants to conditions—will control
most potential confounding.

Question 20

A good research design follows these five additional procedures:

Answer 20

1. A clearly stated research hypothesis concerning predicted causal effects of one variable on
   another
2. At least two levels of the independent variable
3. Random assignment of participants to conditions
4. Specific and systematic procedures for empirically testing the hypothesized causal
   relationships
5. Specific controls to reduce threats to internal validity

Question 21

Whats the difference between discrete and continuous distributions?

Answer 21

Discrete distributions have no numbers between the full numbers (after 4 comes 5)

Continuous distributions have infinite numbers between the numbers (between 4 and 5)

Question 22

What are the characteristics of a normal distribution?

Answer 22

1. it has the shape of a bell
2. is symetrical around the mean µ
3. Many variables that vary by chance are approximately normally distributed, with mean and standard deviation free to vary
4. µ and and σ deviation are variable

Question 23

What is the standard normal distribution?

Answer 23

A normal distribution with μ = 0 and σ = 1. Any normal distribution can be transformed to a standard normal distribution:
z=(x-µ)/σ