CDS 501 Week 5

• evaluated as separate individuals, not a group  
• detailed analysis of behavior under controlled conditions 
• series of measurements of DV over time 
• each participant receives all level of IV

all persons of interest for a study

people who actually participate in a study

numbers that come from observation of a population

numbers that come from observation of a sample

• good match for characteristics of the general population
• select subjects that represent everyone so that we can make inferences from the study (inductive reasoning)

all members have an equal chance of being selected

• a drawing 
• equal chance of being assigned to either control or experimental group
• even distribution of participants within the sample

• establish sampling interval 
• select subjects according to the number representing the sampling interval

• divide possible subjects into groups 
• select people from subgroup
• make a sample that represents the diversity in the population being studied

• from a large population 
• usually regional studies 
• ID predefined groups from a wide area range (all kindergarten classes in the district)

• get perspective of individual, unique experiences 
• select participants considered to be the best source for particular issue 
• goal not to generalize findings to larger population

• choose subjects from available pool of people 
• align as closely as possible 
• major source of bias

• ability of study to detect significant treatment effects when present 
• too low and the treatment is ineffective

attempt to match two groups on as many extraneous variables as possible

• does a test measure what it says it will measure?
• the test is made for a purpose

• are the results consistent and repeatable
• the test is consistent under certain circumstances
• test may be reliable only when child is attentive

person's judgement on how well a test appears to accomplish its purpose

established by a panel of experts in particular content area

comparing a new test/measure to an existing one that serves as a standard of comparison

• same subjects take new test and established test close together or at the same time
• *type of criterion validity

• predict how participants will perform when  reevaluated at a later time 
• did scores on the old test successfully predict their future performance?

• patterns or relationships among test items 
• relationship between test items and external patterns of comparison 
• how we expect kids to respond to test items
• ex) kids know 50 words by 18 months

• two different examiners use procedure to test the same group of participants around the same  time 
• the two different examiners should achieve similar results

subjects tested by same examiner two times and the results are similar

occurs after researchers have developed and administered two equivalent forms of a measure

• test items are split in two equivalent forms 
• scores from each form are compared

*you don't have to look at the whole test, just look at half to compare the changes over time  

• how we control for interference of extraneous variable 
• degree to which relationship between IV and DV is observed without the influence of extraneous variables
• Did the experimental treatments make a difference in this instance

• *most common weakness in research studies are related to internal validity 
• *affects the researcher's ability to show cause and effect between IV and DV

• how comfortable we are making inferences about the general population from the results of our study 
• can used to predict behavior 

*it can't apply to everyone, that's why we replicate it in different settings with different populations

• how the setting interacts with the IV and changes the DV 
• commonly known as Hawthorne effect: subjects behavior changes because they know they are in a study

• how various parts of multiple treatments affect the DV 
• likely when more than one treatment given to same group of subjects

• how externally valid qualitative findings are
• ability to apply results of research to multiple, similar contexts

• protects internal and external validity 
• subjects selected using preset criteria 
• subjects receive exactly the same instruction 
• cannot modify design

are the subjects representative of the population the research is examining?

how reactive pretest may interact with the IV and affect the DV