As in our experiments on the influence of direct and indirect speech on source and item memory, we will present participants with either an indirect or a direct speech version of a story. This story is about four protagonists (two males, two females) who go to a pub together and talk about 12 different topics. Each protagonist makes a statement with respect to each topic, so the story contains 4x12=48 statements that are presented either in indirect or direct speech. Because of the length of the story, we will present the story in four segments. The shortest segment contains 234 words, the longest 350. Reading times for each part will be recorded and used to exclude extremely fast participants (see Analysis). After participants read the whole story, they will be presented with two blocks of questions. In one block, participants have to indicate for 24 statements who the protagonist was that made that statement in the story. Answer options are the names of the four protagonists. We selected two statements of all 12 subjects that are discussed in the story and we made sure that we included 6 statements of each protagonists. Statements will be presented one at a time and in random order. The answers given by the participants will be recorded and used to calculate a score for the total number of correct answers (correct name) and a score for the total number of correct answers based on gender of the protagonist. These scores represent source memory. Our main interest is in protagonist name, as this is the most direct measure of source. The other block consists of 24 statements that either represent information mentioned in the story or not (e.g., One of the colleagues asked the waiter for the menu). Half of the statements consists of information mentioned in the story and half of them not. Participants have to indicate if the information in the statement was mentioned in the story ('yes') or not ('no'). They respond by clicking on one of two answer options. Statements will be presented one at a time and in random order. The answers given by the participants will be recorded and used to calculate a score for the total number of correct answers. This score represents memory for the textbase of the story. The order in which the two blocks of questions will appear is also randomized. We will subsequently ask participants to take their best guess as to what the purpose of this task was. In addition we will ask them to indicate whether there were any noise or distractions while performing the experiment and what device they used to perfrom the task. Also, we will ask them to state their gender, age, level of education, native language, and the country where they live. We expect this experiment to take about 15-20 minutes to complete. Participants will receive $1.5 for their participation. As in our previous experiments, participants will be recruited trough Amazon Mechanical Turk (http://www.mturk.com). The experiment will be presented online in the Qualtrics survey research suite (http://www.qualtrics.com). According to a power analysis (https://osf.io/g5ehj/), we need at least 60 valid participants per condition to reach a power of .80 for the interaction, conservatively assuming that source and content memory are correlated .3. However, it is possible that first having answered the content questions (as will be the case for half our subjects) will diminish the speech effect. We will therefore aim for a larger number of valid subjects. We will recruit 250 participants who will be randomly assigned to either the indirect or the direct speech condition. The data will be analyzed by a 2 (indirect vs. direct speech) X 2 (source vs. content memory) mixed ANOVA, with speech as a between-subjects and memory as a within-subjects factor. These analyses will be followed up with frequentist and Bayesian t-tests.