Association for Behavior Analysis International

The Association for Behavior Analysis International® (ABAI) is a nonprofit membership organization with the mission to contribute to the well-being of society by developing, enhancing, and supporting the growth and vitality of the science of behavior analysis through research, education, and practice.

Search

34th Annual Convention; Chicago, IL; 2008

Event Details


Previous Page

 

Poster Session #208
#210 Poster Session (EAB)
Sunday, May 25, 2008
12:00 PM–1:30 PM
South Exhibit Hall
95. Sequential Effects During Assessment of Delay Discounting: Ascending vs. Descending Amounts of Immediately Available Cash.
Area: EAB; Domain: Basic Research
ELIAS ROBLES (Arizona State University), Perla A. Vargas (Arizona State University)
Abstract: Procedural variants in estimating delay discounting (DD) have been shown to yield significant within-subject differences in estimated degree of delay discounting as well as variations in the patterns of choice. This study evaluated, within-subjects, the effect of order of presentation of the immediate rewards in a delay- discounting task. Participants (n = 75) were assessed with two computerized DD tasks. The tasks differed on the order of presentation of the immediately available rewards. Greater mean area under the curve (AUC) was estimated when the descending sequence was used. The most frequent location for the longest reaction time in each delay value series was the indifference point trial. Subjects tended to switch early when the amount of the immediately available reward decreased with every trial, and to switch later when the amount of the immediately available reward increased from trial to trial, thus yielding significantly different amounts of estimated delay discounting. It is possible, that the observed differences may reflect framing effects since the differences are very consistent and do not appear to be related to a reduction in the number of trials, task duration, or subject self-reported level of difficulty or interest.
 
96. Measuring the Auditory Abilities of Brushtail Possums (Trichosurus vulpecula).
Area: EAB; Domain: Basic Research
MIZUHO OSUGI (The University of Waikato, New Zealand), William Temple (The University of Waikato, New Zealand), Therese Mary Foster (The University of Waikato, New Zealand)
Abstract: Brushtail possums (Trichosurus vulpecula) are omnivores and are a major pest both in indigenous forests and to agriculture in New Zealand, however, little is known of their sensory abilities. There is only one behavioral study examining their auditory ability. This used a two-response conditional discrimination procedure in which the response was a lever press and correct detection of the presence or absence of the tone gained access to food. This study examined possums’ ability to detect 880 Hz tones at various intensities, using a modified tracking procedure, and found that they could just detect this frequency at 36 dB (A). The present study used the same procedure to investigate their ability to detect tones of higher and lower frequencies than 880 Hz. Eight possums were initially trained to perform the conditioned discrimination with an 880 Hz tone at 80 dB (A). Tones of 2000, 4000, 6000 and 10,000 Hz were then used. Initial data shows very similar performance as intensity decreased for all these tones with detection thresholds at around 30 to 40 dB (A) for which no consistent trend across tones. Further results of the study will be presented and discussed.
 
97. EAHB SIG 2008 Student Paper Competition Winner: Establishing Concepts of Inferential Statistics and Hypothesis Decision Making Through Contextually-Controlled Equivalence Classes.
Area: EAB; Domain: Basic Research
DANIEL MARK FIENUP (Illinois State University), Thomas S. Critchfield (Illinois State University)
Abstract: This study examines the acquisition of concepts related to inferential statistics and hypothesis decision making. Stimulus equivalence methodology was employed through three instructional lessons. The first lesson taught participants basic responding to inferential statistics information. The second lesson taught participants to make decisions about hypotheses (e.g., reject, fail to reject) based on one’s predictions and direction of results. The third lesson taught participants the conditional influence of inferential statistics information over decisions regarding the scientific and null hypotheses. Following all training, extended test batteries were given to assess whether the contextual control, established in Lesson 3, transferred to other stimuli. In total, participants were explicitly taught 40 relations which resulted in the emergence of 144 relations. This study demonstrates the efficiency of employing stimulus equivalence methodology in the acquisition of higher level concept learning.
 
98. Undermatching and Overmatching in a Choice Situation with Variable Changeover Requirements.
Area: EAB; Domain: Basic Research
ÁNGEL JIMÉNEZ (University of Guadelajara), Carlos F. Aparicio (University of Guadalajara)
Abstract: When choice reaches steady state, sensitivity to reinforcement increases with increasing changeover requirements. This study assessed with rats the generality of this finding in a situation where choice did not reach stability. Two concurrent schedules of variable interval arranged reinforcers on average every 12 seconds contingent to the behavior of lever pressing. In order to simulate a variable environment, every session programmed five components that required 1, 4, 8, 16, or 32 responses to switch from one lever to the other. In sessions that ended after the delivery of 50 reinforcers, the components occurred in random order and without replacement. Every component delivered 10 reinforcers and ended with a 1 min blackout. One of seven reinforcer ratios (27:1, 9:1, 3:1, 1:1, 1:3, 1:9, and 1:27) were used every 21 sessions. Response and visit time distributions adjusted to the changes in reinforcer distributions. The slope of the generalized matching equation estimated with responses increased with increasing changeover requirements. Sensitivity estimated with visit times showed a bimodal function. The implications that these findings have for a molar model of choice are discussed.
 
99. Shot Selection as Operant Choice: Do Individual NBA Players Obey the Matching Law?
Area: EAB; Domain: Applied Research
JENNIFER L. HITT (Illinois State University), Larry Alferink (Illinois State University), Thomas S. Critchfield (Illinois State University)
Abstract: Several previous studies have found that the ratio of two-point to three-point shots taken by basketball players tracks the ratio of shots made in a way that is consistent with the matching law. All previous analyses, however, have aggregated data from several individuals, whereas the matching law is a statement about how INDIVIDUAL behavior relates to its consequences. We describe attempts to analyze archival data of individual National Basketball Association players using the matching law and show that questions may be raised about the extent to which the "reinforcement-matching" that has been described at the aggregate level accurately portrays shot selection by individual players.
 
100. Delay Discounting of Money, Gift Certificate for Food and Food.
Area: EAB; Domain: Basic Research
ANA AMELIA L. BAUMANN NEVES (Utah State University), Wesley P. Thomas (Utah State University), Amy Odum (Utah State University)
Abstract: Delay discounting determines how the value of the outcome is affected by the delay to its receipt. Research to date has shown that consumable outcomes (e.g., alcohol) are discounted more steeply than the non-consumable outcome money. We have been investigating the generality of this effect by comparing discounting of food (directly consumed), and money (not consumed but exchanged for other things) with other commodities. Our previous study showed that food and gasoline are both discounted more steeply than money. In the present study, we investigated whether the discounting curves would differ for food, gift certificates for food, and money. Human participants completed a computer-based titration procedure to estimate the degree of discounting for hypothetical outcomes at seven different delays. Each participant completed the assessment for all three outcomes separately. The data indicate that the degree of discounting for food is larger than the degree of discounting for gift certificates for food. The degree of discounting for gift certificates was similar to that of money. These results suggest that discounting may be affected by the degree of consumability and exchangeability of the outcomes.
 
101. Choice, Preference and Self-Determination in Mentally Retarded Adults.
Area: EAB; Domain: Basic Research
A. CELSO GOYOS (Universidade Federal de São Carlos), Giovana Escobal (Universidade Federal de São Carlos), Rubens Andreolli (Universidade Federal de São Carlos)
Abstract: Choice was investigated with and without environmental work support. Also, work performance was assessed as a function of conditions chosen. Four mentally retarded adults learned a work task with and without environmental work support. The environmental support displayed containers to place its different components and was designed to provide immediate feedback, to increase or maintain the rate of work response and to prevent errors during the task routine. Following initial training the subjects worked individually either under multiple schedules or under concurrent-chain schedules according to a multiple-element design. When the multiple schedules condition was on, the components were either presence or absence of environment support, quasi-randomly distributed. In the concurrent-chain schedules condition it was used a FR-1, on the first link, and either one of the presence-absence of environmental support condition, on the second link. Results showed that when choice opportunities were given, the work support condition was chosen more often and time spent to task completion and average number of errors performing the task decreased. The results also suggested that the environmental work support yielded greater control over work and the opportunity to make choices engendered more motivation and work independence in the participants.
 
102. Extinction Following Schedules of Continuous, Intermittent and Non-Contingent Reinforcement.
Area: EAB; Domain: Basic Research
RYAN J. BRACKNEY (University of North Texas), Erik Arntzen (Akershus University College )
Abstract: It has been suggested that extinction proceeds more rapidly if preceded by a history of continuous reinforcement than if preceded by a history of intermittent reinforcement. Laboratory investigations of this phenomenon, however, have yielded contrary findings. The data from some studies have shown an interpolation effect— interpolating a history of continuous reinforcement has produced more rapid extinction. Other studies, however, have failed to document such effects and still others have actually shown that the pace of extinction was retarded relative to conditions in which there was no interpolation of a CRF history. The current study sought to investigate this issue by (1) establishing robust key pecking in pigeons, (2) maintaining it on a moderate-value intermittent schedule of reinforcement, and (3) measuring the extent of and delay to extinction following interpolation of CRF or non-contingent food delivery conditions. The results from this study have the potential to inform the use of extinction with high frequency problem behavior.
 
103. Immediate Post-Session Feeding Reduces Progressive Ratio Breakpoints.
Area: EAB; Domain: Basic Research
JOHN R. SMETHELLS (Central Michigan University), Jennifer Andrews (Central Michigan University), Andrew T. Fox (Central Michigan University), Mark P. Reilly (Central Michigan University)
Abstract: Two experiments investigated the effects of post-session feeding on progressive ratio break points in rats. Across conditions of the experiments, post-session feeding was either provided immediately following the session or delayed t minutes following the session. Break points, the last ratio completed prior to a 5 m pause in responding, were lower during immediate post-session feeding, both when deprivation levels were controlled with regard to body weight (Experiment 1) or time elapsed since the previous meal (Experiment 2). These findings support those of Bacotti (1976) who showed that post-session feeding decreased variable-interval response rates. Post- session feeding times should be reported and carefully controlled to ensure the consistency of a food reinforcer’s efficacy.
 
104. Controlling Eating Behavior in Rats with Periods of Access and No-Access to Food over 24 Hours.
Area: EAB; Domain: Basic Research
FELIPE RESENDIZ DIAZ (Universidad Nacional Autonoma de Mexico of Mexico), Carlos A. Bruner (Universidad Nacional Autonoma de Mexico of Mexico)
Abstract: In a prior study we analyzed the daily eating patterns of rats with unrestricted access to food and water. Using the method of survivor analysis, eating was described as bouts of a given duration and periods between bouts. The present study investigated the effects of deliberately controlling the durations of both the eating periods and the inter eating periods over 24 hours. Given that the eating periods were signaled, these manipulations can be seen as a Multiple Fixed Ratio 1 EXT schedule of reinforcement of unusually long component durations. The duration of the EXT components was either 45, 180 or 720 min and the duration of the reinforcement component was either 2.5, 10 or 40 min. The dependent variable was the rate of lever pressing for food during the reinforcement components. Three rats were assigned to each combination of the variables. Combinations of 80 minreplicated the normal daily-eating time seen in the previous study. The results showed that 80-minute variable combinations that deviated from the normal eating pattern controlled either higher or lower response rates. Also, that combinations that produced a scarcity of food yielded higher rates than combinations that yielded an abundance of food. The best prediction of responding-for-food rates was component- duration interactions in the form of relative-time effects.
 
105. Behavioral Economics: Hens’ Demand for Food Over Different Length Sessions.
Area: EAB; Domain: Basic Research
THERESE MARY FOSTER (University of Waikato, New Zealand), Jennifer M. Kinloch (University of Waikato, New Zealand), William Temple (University of Waikato, New Zealand)
Abstract: Controlling the degree to which an animal’s food requirement is provided in an experimental session can involve varying session length and such covariation may confound interpretation of the results of such studies. In this study hens responded under fixed ratio schedules and session lengths of 2 hr, 1 hr, 40 min, and 10 min. In each condition the fixed ratio requirement started at 1 and was doubled each session until a hen received no reinforcers in a session. The fixed ratio was then set at 20 for that hen for the remainder of the condition. The condition ended when all hens had completed that fixed ratio series. There were at least two conditions with each session length. Comparison of the data under the fixed ratios for the same length periods of the different length sessions showed that response patterns in shorter sessions were similar to those over the same period of time of a longer session. A demand analysis also showed similar functions over similar periods of all length sessions but also showed that demand became increasingly more inelastic over the smaller fixed ratios as data from longer periods of the session were included.
 
106. The Dynamics of Stimulus Bias.
Area: EAB; Domain: Basic Research
DANIEL LEE HUTCHISON (Jacksonville State University), William L. Palya (Jacksonville State University)
Abstract: Pigeons were exposed to a multiple VI / extinction procedure. Different groups of pigeons were trained with VI stimuli of either red, green, or blue. The extinction portion of the multiple schedule was in the presence of either a white key light or darkness. This resulted in 6 different groups. A seventh group received a white S+ and a dark S-. Phase 2 implemented a VI with 6 s each of 11 different hues. The key was continuously illuminated and the stimuli were randomly presented and independent of the reinforcement contingency. In general prior exposure to a stimulus as an S+ resulted in higher rates to that stimulus for a short period. The data showed that the bias tended to be stronger in the condition with white as S- rather than a dark key. Responding to each color became more similar with increasing experience.
 
107. Transfer of "Good" and "Bad" Functions Within Stimulus Equivalence Classes.
Area: EAB; Domain: Basic Research
JESSICA A. MADRIGAL-BAUGUSS (University of North Texas), Sigrid S. Glenn (University of North Texas)
Abstract: This study compared results of two between-subject experiments that tested transfer of function in stimulus equivalence-based classes in a task dissimilar to (in Experiment 1) and similar to (in Experiment 2) the task that trained the original conditional discriminations among class members. Seven undergraduate and 4 graduate students from UNT participated in return for monetary compensation. The first two phases in each experiment were identical. Phase 1 presented a conditional discrimination task to establish three 4-member stimulus equivalence classes. Phase 2 presented a successive discrimination task that awarded money on a VR20 schedule for responding to a stimulus of equivalence class 1 and subtracted money on a FR5 schedule for responding to a stimulus of equivalence class 2. In Experiment 1, Phase 3 involved a simple simultaneous discrimination task to assess preference between classes. In Experiment 2, Phase 3 involved a successive discrimination task to assess response rates to each class. Participants in Experiment 2 demonstrated consistent transfer of function, whereas participants in Experiment 1 did not. Results are discussed in terms of how task similarity may function as a type of contextual control when there is limited experience with the task.
 
108. Reducing Smoking in College Students Using a Percentile Schedule with Fixed-Value Voucher Reinforcement: A Feasibility Study.
Area: EAB; Domain: Basic Research
J. PHILIP ERB (James Madison University), Kristin A. Kiel (James Madison University), Brantley P. Jarvis (James Madison University), Jessica Greta Marie Irons (James Madison University)
Abstract: Contingency management (CM) programs have been demonstrated effective for reducing smoking in a number of populations including college students. However, within each of those populations, there is a subset of individuals that fail to demonstrate the abstinence criteria specified by the contingency and thus fail to earn the programmed reinforcers and ultimately experience success. One reason for this may be that that abstinence criterion is too difficult to achieve. Percentile schedules alleviate this problem by basing the criterion level of behavior required for each reinforcer on recent samples of that given individual’s behavior, thereby increasing the probability of the individual contacting the contingency. Research has shown that percentile schedules are effective in reducing smoking when used in combination with escalating-value reinforcers. The current study was designed to evaluate the effectiveness of a CM program using a percentile schedule with fixed-value reinforcers in a college student population. To the extent that participants’ smoking is reduced, the effectiveness of this procedure in this context is supported. In addition, to the extent fixed as opposed to escalating values of reinforcement are effective, these data have important implications for increasing the cost-effectiveness of CM procedures.
 
109. DRL Responding Across a Changing Photoperiod in Migratory White-crowned Sparrows.
Area: EAB; Domain: Basic Research
ELLIOTT M. PALETZ (University of Wisconsin, Madison), William H. Obermeyer (University of Wisconsin, Madison), Ruth M. Benca (University of Wisconsin, Madison)
Abstract: In the wild, migratory birds increase their locomotor activity during biannual migrations. Smaller species, like the sparrow, are easy to maintain in the laboratory, and if the photoperiod is repeatedly adjusted to simulate ecological conditions, the birds will engage in stationary flight within their cages, roughly at times when they would normally migrate in the wild. In previous studies using fixed and progressive ratios, white-crowned sparrows increased their response rates in close temporal proximity to observed increases in home-cage activity. In the present experiment, white-crowned sparrows (Z.l. gambelii) were exposed to a differential reinforcement of low rate (DRL 20-s) schedule of reinforcement to determine if generalized increases in activity perturb relatively low baseline rates of responding. The birds were captured in California, housed one per cage in the laboratory, and provided with ad libitum access to food and water. After several months, the experiment began. Three hours prior to each session, food was removed from the home cage. During times when the birds became increasingly active, they increased their response rates despite concomitant decrements in reinforcement. Results are discussed within the context of endogenous pharmacological effects on behavior as well as difficulties in maintaining establishing operations within a migratory species.
 
110. Rapid Acquisition of Bias in Signal Detection.
Area: EAB; Domain: Basic Research
BLAKE A. HUTSELL (Southern Illinois University at Carbondale), Eric A. Jacobs (Southern Illinois University at Carbondale)
Abstract: Four rats were trained to discriminate two sample stimuli (2 or 8 s houselight presentations). The arranged reinforcer ratio for correct responses varied from session to session according to a 31-step pseudorandom binary sequence. Data were analyzed according to the detection model of Davison and Tustin (1978). Analyses showed an overall asymmetry in performance across the two sample stimuli. Both sensitivity to reinforcement and discriminability were higher for each subject in the presence of the shorter duration stimulus. Bias (log b) tracked the random changes in obtained reinforcer ratio and discriminability (log d) was relatively constant across sessions. The detection model was extended to quantify the contribution of current and previous sessions' reinforcer ratio to current session sensitivity to reinforcement. Similar to findings using concurrent schedules, sensitivity to reinforcement was highest in the current session, however effects of previous sessions were evident.
 
111. Size Transposition with Two Training Pairs in Pigeons.
Area: EAB; Domain: Basic Research
NAOYA KUBO (Komazawa University), Koichi Ono (Komazawa University)
Abstract: The purpose of this experiment was to determine whether size transposition would take place when pigeons were trained by two pairs of stimuli. Pigeons discriminated two pairs of triangles: 1-3, 4-6 (1 was smallest stimulus, 6 was largest ). S+ stimuli for 2 pigeons were smaller triangle; S+ for other pigeons were larger triangle. In Test 1, four pairs were presented; two pairs of test stimuli with same history(1-4,3-6), a pair of novel stimuli (2-5), and a pair of combined stimulus with S+ history and novel stimulus (2-4 or 3-5). In Test 2, to examine whether the functions acquired in training were replicated by the stimuli pairs including the stimulus used in training, four pairs were presented; 1–6, 2–6, 1–5 and 3–5 or 2–4. As a result of the experiment, pigeons nearly responded to relative rather than absolute size showing transposition partially. And the function of the training stimuli were maintained in Test 2.
 
112. Delay Discounting Research with Problematic Sexual Behavior in Human Subjects.
Area: EAB; Domain: Applied Research
JESSE M. CROSBY (Utah State University), Mike P. Twohig (Utah State University), Timothy A. Shahan (Utah State University), Jeffrey W. Allen (Utah State University)
Abstract: Delay discounting refers to decreases in the value of a reinforcer as a function of the temporal delay of that reinforcer. This decrease in value manifests as a decrease in the effectiveness of the reinforcer on behavior. It can be measured for an individual and described as a hyperbolic function—the steepness of which indicates the level of impulsivity in the individual. Research with delay discounting has identified significant differences in impulsivity between experimental and control groups in substance use and gambling populations. Problematic sexual behavior (i.e., addictive viewing of pornography or compulsive masturbation) has been conceptualized as addictive, compulsive, or impulsive behavior. This study will use delay discounting as a measure of impulsivity with this population and report comparisons with controls and correlations with other measures related to problematic sexual behavior. Data will be collected from an undergraduate university population in an anonymous survey format.
 
113. Differences in Sustained Operant Variability Levels.
Area: EAB; Domain: Basic Research
MICHAEL YOUNG (Southern Illinois University at Carbondale), Eric A. Jacobs (Southern Illinois University at Carbondale)
Abstract: Contingencies calling for variability at levels both above and below the baseline level elicited by the task were introduced. Participants pressed four-key responses on a computer keyboard in one of four contingency conditions: a high-variability lag 50 contingency (L50-0), a lag 50 contingency combined with a within sequence variability requirement (L50-2), a low-variability lag 1contingency (L1-0) and a lag 1 contingency combined with a within sequence requirement (L1-2). Between-sequence entropy and within-sequence entropy were measured. Within-sequence entropy was increased relative to baseline in the L50-2 and L1-2 conditions, between-sequence entropy increased in the L50-0 and L50-2 conditions and between-sequence entropy decreased in the both L1-0 and L1-2 conditions. Decreases in between-sequence entropy and increases in within-sequence entropy during the contingency were partially sustained in extinction, while increases in between-sequence variability were not sustained. Both types of variability contingency seemed to bring the type of variability they addressed under the control of the contingency while they were in effect, but they had differential effects on sustained variability in extinction.
 
114. Social Learning in Flocks of Pigeons: Effects of Density Population.
Area: EAB; Domain: Basic Research
ROSALVA CABRERA (FES Iztacala - National University of Mexico), Abel Javier Zamora (FES Iztacala-National University of Mexico), Martha Elisa Lopez (FES Iztacala-National University of Mexico)
Abstract: In Experiment 1, the acquisition of novel responses by observers was evaluated in flocks of naïve pigeons which were composed by few or a lot of members; in Experiment 2, the acquisition of novel responses by observers was evaluated when they were exposed to one or many trained demonstrators. In both experiments were evaluated the strategies developed by the members of flocks. In the experimental situation, the trained demonstrator(s) perform the response of piercing seals to obtain mixed grain to flocks of naïve subjects. In Experiment 1, the number of observers by group was varied (4, 8, or 12 observers) by flock. In Experiment 2, the number of demonstrators presented to a flock was varied (1, 2, or 4 demonstrators). The data show that the acquisition of the response by observers was an inverse function of the number of observers in the flock (Experiment 1) and of the number of trained demonstrators modeling the response. In both experiments, the index of scrounging and competition in the flocks increased with the number of members in each flock. On one hand, a fine analysis of data show that the observation of response-food relation by the observers is crucial to learning by observation. On the other hand, the strategies developed by members of flocks can interfere with the performance of a modeled response.
 
115. Right Versus Wrong Feedback in Second-Order Matching-to-Sample I: The Case of Instructional Stimuli.
Area: EAB; Domain: Basic Research
MARIO SERRANO (Universidad de Guadaljara), Emilio Ribes Iñesta (Universidad de Guadalajara), Gustavo Garcia (Universidad FrancoMexicana-Satelite)
Abstract: Three groups of college students were exposed to an instructional second-order matching-to-sample task and intramodal, extramodal, and extradimensional transfer tests. Instructional second-order stimuli consisted in pairs of geometric shapes that visually modeled the ongoing matching relation in each trial. For different groups, only right, only wrong or both right and wrong matching responses produced the correlated feedback during the training. Percentage of correct responses in both training and transfer tests were higher for the participants exposed to the procedure in which only wrong matching responses produced feedback, while no difference between the other two groups was observed. Results are discussed in relation to previous experiments in which feedback for both right and wrong matching responses was scheduled in either intermittent or delayed fashion.
 
116. The Schedule-Induced Drinking Procedure Does Not Generate Excessive Water Intake.
Area: EAB; Domain: Basic Research
ALICIA ROCA (National University of Mexico), Carlos A. Bruner (National University of Mexico)
Abstract: The schedule-induced drinking (SID) procedure involves depriving rats of food and delivering food at intervals. Two experiments were conducted to determine the role of both operations in the occurrence of SID. In Experiment 1, three rats at 80% of their ad lib weight lived in the experimental chambers. During the first condition, three 1-hour SID sessions were intruded into each 24-hour period. On each session food pellets were delivered on a 180-s fixed-time schedule in order to complete 1, 3, or 8 gs. During the second and the third conditions, the food was delivered altogether at the beginning of each session or at the beginning of each 24- hour period. Under spaced food delivery, drinking occurred within the sessions, and was proportional to meal size. Water intake remained constant during the three conditions. In Experiment 2, the fixed-time schedule was reinstated and rats were kept at 100% or 70% of their ad lib weight in successive conditions. Increasing food deprivation controlled increases in water intake during the sessions, accompanied by decreases outside the sessions. These data show that spaced food delivery and food deprivation do not generate excessive drinking, but rather redistribute total daily water intake.
 
 

BACK TO THE TOP

 

Back to Top
ValidatorError
  
Modifed by Eddie Soh
DONATE