PSYCH STATS

• Albert Einstein: 'It's not that I'm so smart, it's just that I stay with problems longer.'
• Interview


  The face-to-face question-and-answer process between a researcher (interviewer) and participant (interviewee)
• Case study


  The in-depth gathering of data about one individual
• Researcher aims to study something – animals, humans, or objects. The totality or entirety of that which one wants to study is referred to as the population
• If the defined population of a study is small enough to enable the researcher to gather data from the whole population, we call it a census
• The good news is – we can still study the population even if we only get a small portion of it so long as that portion is chosen correctly. A portion of the population is called a sample
• Good sample

  • Representativeness - the sample is very much the same as the population
  • Sufficiency - the sample consists of an adequate number of entities to be considered a representative of the population
• Random sampling


  Probability sampling where each member of the population has a chance or possibility of being selected to participate in the study
• Simple random sampling


  The sample is chosen based on pure chance, often done by lottery, draw lots, or fishbowl technique
• Systematic random sampling


  The sample is chosen based on a system like all the 10th households in the city or every 50th store customer
• Stratified random sampling


  Ideal for populations made up of stratum or groups whose population is significantly different from each other, duplicates the proportion of the groups in a population
• Cluster sampling


  Commonly used when the population is naturally composed of strata or clusters (small subsets that relatively mimics the characteristic of the whole population)
• Non-random sampling

  Non-probability sampling where samples are chosen in some specific manner by the researcher which means that it is relatively and inherently biased compared to random sampling
• Convenience sampling


  Participants or objects in the sample are selected based on their availability to the researcher
• Purposive sampling


  Participants are chosen based on the specific description imposed by the research topic
• Quota sampling


  Very much like the stratified sampling, however, the quota or specified proportion is arbitrarily set for each group. Then, sample selection is done by convenience sampling
• Snowball or referral sampling


  When the topic of interest is rare and finding samples is difficult, a researcher may ask participants to refer them to other members of the population if they know any
• Margin of error


  The level of precision required, often expressed in percentage points (e.g., ±2%)
• Confidence level


  The amount of uncertainty associated with an estimate, the chance that the confidence interval will contain the true value
• Data


  Factual values collated for empirical analysis
• Qualitative data


  • Describes an organism or a number that stands for characteristic, attribute, quality, or category
  • Answers questions pertaining to what, where or when
• Nominal data


  Concepts of relative position like better than or greater than cannot be assumed
• Ordinal data


  There is a concept of sequence and arrangement but the differences between categories cannot be measured outright nor assumed to be the same
• Quantitative data


  • Numerical type of data that gives an idea of the amount, measure, quantity, or magnitude of anything
  • Answers "how much" questions
• Interval data


  Quantitative information or number that implies order and equal magnitude between values but has no absolute zero
• Ratio data


  • Data that results from quantitative measurements like weight, height, allowance per day
  • The differences between values are the same and absolute zero exists
• Interval-ratio data


  • Statistical techniques for interval and ratio are the same
  • Zero is not considered an absence of something or the non-existence of the variable
• Score


  Numerical data derived from measurements using machines or tests
• Proportion


  Data derived from counting objects, participants, or variables from a given population
• Univariate analysis


  Data analysis with only one variable, to describe the data and find patterns
• Bivariate statistics


  Statistics analyzing the relationship between two variables
• Multivariate statistics


  Statistics analyzing the relationship of more than three variables
• Bivariate statistics


  • t-test
  • z-test
  • r (correlation)
• Multivariate statistics


  • Analysis of Variance (ANOVA)
  • Regression (predicting based on the contribution of different variables)
• Within-groups


  Only one group of respondents, participants are measured repeatedly in different conditions or matched
• Between-groups


  Two or more groups of respondents, participants are measured only once and their scores are considered in one group alone to be compared to another group
• Statistic


  A value computed from a sample
• Parameter


  A value computed from or claimed about the population
• Parametric vs Non-parametric statistics

  • Parametric: Normally distributed data, same variance, interval-ratio level of measurement
  • Non-parametric: May deviate from normal distribution, variance may not be the same, nominal or ordinal level of measurement
• Parametric vs Non-parametric statistical techniques

  • Parametric: Mean, Pearson correlation, t-test, ANOVA
  • Non-parametric: Median/mode, Spearman correlation, Mann-Whitney, Kruskal-Wallis, Wilcoxon