• Reliability : the degree to which the results of a study can be replicated under similar conditions
  • Operational definition : the definition of an abstract concept (such as intelligence) used by a researcher to measure or manipulate the concept in a research study
  • Qualitative data: non­numerical participant responses
  • Quantitate data: numerical data
  • What are the 4 primary scales of measurement for dependent variables?
  1. Nominal: scale of data measurement that involves non­ordered categorical responses o Simplest scale

o Examples: measurements of mood, major, gender o Qualitative data

  1. Ordinal: scale of data measurement that involves ordered categorical responses o Can be ordered from highest to lowest
  2. Not at all, a little, fairly, very
  • Not necessarily equally spaced ii. None, one, 2­5, 5 or more
  • Ordered from smallest to largest
  • None and one is NOT the same difference between one and 2­5

iii.      Rank order of stimuli according to preference: first, second, third, fourth

 The difference between the first and the second choices may be much larger or smaller than the difference between the second and third choices

o Qualitative data

  1. Interval: scale of data that involves numerical responses that are equally spaced o Likert scale: Agree/Disagree with a statement on a scale of 1­5 or 1­7, etc o Quantitative data

o Fahrenheit and Celsius because 20 degrees is NOT twice as cold as 10 degrees

  1. Ratio: scale of data measurement that involves numerical responses where the scores are ratios of each other o Reaction time, distance, time, height, weight o The amount of time to complete a task

 The second participant took twice as long as the first participant o 0 is the lowest unless it’s time because time can be negative o Kelvin does include a 0 value that is the lowest possible value

  • Validity : the study’s accuracy in testing the researcher’s hypothesis o The more values/responses, the more accurate
  • Construct validity : indicates that a survey measure the behavior it is designed to measure o Poor operational definition=LOW construct validity
  • The choice of measurement can also affect the reliability
  • Nonverbal scale : pictorial response categories for participants with low verbal skills
  • Face validity : one the surface, a study or scale appears to be intuitively valid o If you want to know about someone’s mood, asking respondents about their mood seems as if it will measure what you want to measure
  • Inter­rater reliability: a measure of the degree to which different observers rate behaviors in similar ways
    • This is most important in naturalistic observation because there are multiple observers observing at the same time
    • High inter­rater reliability means there is a high rate of agreement in the way the different observers are recording behaviors
  • Presence/absence variables: a variable with a level that involves the treatment and a level that does not involve the treatment o Experimental group and control group
  • Bivalent independent variable: an IV with two levels
  • Type variable: variable that involves a manipulation of types of a treatment
  • Amount variable: variable that includes levels with a different amount of the treatment
  • Multivalent variable: an IV that includes 3 or more levels
  • Quasi­independent variables: allows comparison of groups of participants without manipulation o AKA the subject variable
    • Gender, age, ethnicity, and personality type
  • Internal validity: the degree to which a study provides causal information about behavior o The best way to construct a good experiment is to control for sources of bias that can make the causal relationship less clear

 These kinds of bias are known as confounding variables

  • Confounding variable: an extraneous factor present in a study that may affect the results
  • Experimenter Bias: when a researcher treats groups differently (often unknowingly) based on knowledge of the hypothesis
  • Testing effects: when participants are tested more than once in a study with early testing affecting later testing o More likely to occur in within subject studies

o Occurs when participants are more fatigued the second trial or when they improve overtime

  • Single blind design : used to hide the group assignment from the participants
  • Double­blind design : keeps the knowledge of the group assignments from BOTH the participants and the researcher
  • Within subject variable: each participant experiences ALL levels
  • Between subject variable: each participant experiences only ONE level
  • Counterbalance: different participants are assigned in equal numbers to the different orders of the conditions
  • Regression toward the mean: occurs when participants obtain an extreme score (high or low) on a questionnaire or task at one testing session but regress toward their mean score at another testing session o Lowers the validity of the test
    • Best way to minimize these effects is to use random assignment
  • External validity: degree to which a study measures realistic behaviors and a study’s results can be generalized beyond the study o If participants behave in a research study the way they would in their everyday lives, the study has GOOD external validity
  • Hawthorne Effect: a source of bias that can occur in a study due to participants changing their behavior because they are aware that they are being observes
  • Field experiment: conducted in the participants’ natural environment
  1. Chapter Summary
    • Variable : an attribute, situation, or behavior that can change from individual to individual
    • Dependent variables : measured from individuals and are present in all research studies
    • Confounding variables : may affect the dependent variables in a study but are not of interest to the researchers
    • Independent variables : variables that are manipulated by the researchers in the experiment
    • Quasi­independent variables : present in quasi experiments and experiments; similar to IV in that they involve a comparison of groups, but they are NOT manipulated
    • Which type of research method contains an independent variable? Why?
      • Experiments contain an independent variable
      • They involve the manipulation of a variable that allows for additional control of confounding variables
    • Having good internal validity in a study means reducing bias that can affect the results and providing a good test of the hypothesis
    • Having good external validity means studying behaviors and obtaining results that generalize to individuals beyond the study in their everyday lives
    • Having reliable data in a study means having results that would occur again in a similar situation
    • To control bias, researchers should first identify possible confounding variables in a study and the design the study in the best way to avoid having these variables bias the results
  2. Indicate if it is measured on a nominal, ordinal, interval, or ratio scale:
  3. The amount of money you paid in tuition this year
  • Ratio it cannot go below 0
  1. Your hobbies, from most important to least important to you
  • Ordinal highest to lowest
  1. Whether you plan to go to graduate school (yes or no)
  • Nominal NON­ordered
  1. Your last exam score as a measure of your performance on the exam (as percentage accuracy)
  • Ratio can’t going below 0
  1. How much you are enjoying this class (on a scale of 1­5)
  • Interval
  1. A list of your current career options
  • Nominal NON­ordered
  1. Your class standing (freshman, sophomore, junior, senior)
  • Ordinal ordered
  • For each of the following measures about aspects of beer, match each measure with the correct measurement scale: nominal, ordinal, interval, or ratio
  1. How dark the beer is with choices of lightest, medium light, medium dark, and darkest
  • Ordinal low to high
  1. How long it takes one to identify the type of beer one is drinking
  • Ratio
  1. Which brewery made the beer with choices of Yuengling, Leinenkugel, Blue Moon, Sam Adams, or Coors
  • Nominal
  1. How much someone likes beer being tasted on a scale of 1­5
  • Interval
  1. Questions:
  2. If a score on a midterm is unusually high for a student and far above his or her mean grade in the class, this would represent _______________ as a source of bias in using the midterm score to measure his or her learning in the course
  • Regression toward the mean
  1. A __________________ design is often used to prevent experimenter bias, such that neither the researcher nor the subject is aware of the condition the subject has been assigned to in the study
  • Double blind
  1. Suppose a researcher wants to study work productivity in a factory. Video cameras are installed to see how much time workers spend on tasks during the day. The workers are more productive on the day after the cameras were installed than on the day before. A possible cause of the increase in productivity that would represent a source of bias in the study is ____________
  • The Hawthorn Effect
  1. A researcher is interest in studying face recognition abilities. Subjects are tested in the lab on their recognition of photos of unknown faces presented on a computer screen. If the subjects process the photos of the faces in a way that is different from hos they typically process faces in their daily lives, this study would suffer from low _____________ validity
    • External