Chapters 2, 3, 4, 10, 11, 13, 14, 15 Chapter 2: Experimental research Most common research methods: observations, field studies, surveys, usability studies, interviews, focus groups, controlled experiments o Descriptive investigation: observation, survey, focus groups – construct an accurate description of what is happening (first step- identify interesting phenomena that establish cornerstone of research) o Relational investigation- enable researcher to identify relations between multiple factors. The value of factor X changes as the value of factor Y changes (Discovers connections between multiple variables) o Experimental research- enables the identification of casual relationships between entities or events. It can tell how something happens, and in some cases why (Explore the fundamental casual relations) Research Hypotheses o A precise problem statement that can be directly tested through empirical investigation Hypothesis- smaller, more focused, can be examined by 1 experiment VS Theory- Covers larger scope, requires a sequence of empirical studies (page 23) o Null hypothesis- States that there is no difference between experimental treatments o Alternative hypothesis- always a statement that is mutually exclusive with the null hypothesis H0- There is no difference between the pull-down menu and the pop-up menu in the time spent locating pages H1- There is a difference between the pull-down menu and the popup menu in time spent locating pages Independent variable- type of menu (pull up or pop up) Depend variable- time spent in locating web pages o Goal of experiment is to test the null vs the alternative and decide which one should be accepted and which one should be rejected o Dependent variable- Suggests that the variable is dependent on a participants behavior or the changes in the independent variable (usually outcomes that they need to measure) Efficiency (time to complete task), accuracy (error rate), subjective satisfaction (likert scale usually), ease of learning & retention rate (how quick and easily someone learsn new app or task), physical or cognitive demand (how long without fatigue) o Independent variable- The factors that the researchers are interest in studying or the possible “cause” of the change in the dep variable (usually treatments or conditions that researcher can control) Typical: typing vs speech based, mouse vs joystick or touch pad, pull down vs pop up, font sizes, background colors, age, gender, computer experience, education, mood, lighting, temp, sitting/walking Basics of experimental research (example page 27) o Treatments/conditions- refer to the different research techniques, devices or procedures we want to compare o Units- objects in which we apply the experiment treatments. Normally human subjects with specific characteristics o Assignment method- The way in which the experimental units are assigned different treatments o Randomization- the random assignment of treatments to the experimental units or participants (toss coin, random digital table, software driven randomization). Randomize both assignment and other factors Significance tests o A process in which a null hypothesis is contrasted with an alternative hypothesis to determine the likelihood that the null hypothesis is true Allow us to determine how confident we are that the results observed from the sampling population can be generalized to the entire population o Sampling- Collect a smaller group from the large population and use that smaller group to represent the entire population o Type I error- aka false positive. The mistake of rejecting the null hypothesis when it is true and should not be rejected (example page 32) (worst) (alpha – 0.05) o Type II error-aka false negative. The mistake of not rejecting the null hypothesis when it is false and should be rejected (beta) Limitations of experimental research o Difficult to construct a well-defined and testable hypothesis o Difficult to follow the strict control of factors that may influence the dependent variables o Lab-based experiments not good representation of users typical interaction / Hawthorne effect Chapter 3: Experimental design True Experiments- study involves multiple conditions and the participants are randomly assigned to each condition o Based on at least one testable hypothesis and aims to validate it o Usually at least two conditions or groups o Dependent variables usually measured through quantitative measurments o Should be designed + conducated with goal or removing bias o Should be replicable with dif participant samples, at dif times, at dif locations, by dif experiments Quasi-experiments- study involves multiple groups or measures, but the participants are not randomly assigned to different conditions Non-experiments- only one observation group or only one measure involved Example: There is no difference between the target selection speed when using a mouse, a joystick or a trackball to select icons of different size (s, m, large) o Two independent variables: pointing device + icon size o Three conditions under this: mouse, joystick, trackball o Three more conditions under this: small, medium, large o Total of 9 conditions in this experiment o Dependent variable: speed Decide structure of experiment o How many independent variables do we want to investigate? More than 1? Factorial design Just 1? Basic design o How many different values does each independent variable have? Between-group, within-group or split-plot Between-group design- Each participant is only exposed to ONE condition (Keyboard: one person tries QWERTY, one DVORAK, one Alphabetic) o + Avoids learning effect, only exposed to one condition o + Takes shorter time, fatigue and frustration controlled o – Results are subject to substantial impacts from individual differences o – Larger number of participants needed (4 conditions x 16 people = 64) o Should be used when experiment has simple tasks with limited individual differences, tasks that would be greatly influence by learning effect. Novice & experienced ex, cannot be both Within-group design- Each participant is exposed to all conditions (Keyboard: Three people try all three) o + Requires smaller sample size (16 people) o + Less cost o – Possible impact of learning effects o – Fatigue, takes more time o Should be used when experiment investigages tasks with large individual differences, less susceptible to learning effect, or small participant pool (complicated tasks like reading / writing) Factorial design- Divide experiment groups or conditions into multiple subsets according to the indepdent variables. Split-plot design- One or more indep variables are investigated through betweengroup and other variables are investigated through within-group Interaction effect- The differing effect of one independent variable on the dependent variable, depending on the particular level of another independent variable. Performance may be affected by multiple factors jointly Reliability- Reliable experiments can be replicated by other research teams in other locations and yield results that are consistent, dependable and stable. o Random Errors- can reduce by enlarging observed sample size, closer to observed value (typing speed example) o Systematic errors- aka biases. Always push the observed values in the same direction = too high or too low Caused by measurement instruments- Stop watch reacts slow Caused by experimental procedures- non randomized task conditions = learning effect, not consistent instructions and wording. Should: randomize order, prepare written document with instructions for participant + experimenters, run pilot studies Caused by participants- should: recruit participants carefully (representative of target user pop), create an environment with little stress, reassure participants you are testing interface- not them, reschedule session if participant is tired or nervous Caused by experimenter behavior- Experimenter may untentionally or intentionally influence results (wording, body languaging, late) Caused by environmental factors- Physical: noise, temperature, lighting, humidity. Social: number of people surrounding, relationaship between people and participant