{"id":"86c4dd46-cba8-41ad-ad22-3985b0c2fb47","arxiv_id":"2507.08914","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Nigerian adolescents in the FCT report moderate digital access, frequent reliance on blocking tools, parents as primary support, and a strong demand for better online safety education.","lead":"A survey of 409 Nigerian secondary school students finds that most have moderate internet access, that exposure to inappropriate content and scams is the most common risk, and that blocking and reporting are the leading safety strategies. The study maps an underdocumented adolescent online safety landscape in West Africa and gives parents, educators, and designers concrete local intervention points.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sampling is not a probability sample; the reported ±4.83% margin of error does not support generalization beyond the five selected FCT schools.","rationale":"The paper is a carefully reported descriptive survey with mostly internally consistent frequencies and a plausible thematic analysis. The reader's CONDITIONAL verdict already captures the need for revision. I focus on sampling rather than self-report because self-report is inherent to survey research and the authors took reasonable procedural steps; the sampling design is where the stated precision claim is technically unsupported. The numeric discrepancy between the introduction (59% regular access) and Section 4.2.2 (47% several times a day) is real and should be corrected, and the overbroad 'first study' claim and unreleased instrument are secondary. But the most load-bearing issue is the unearned margin of error: if the schools were not randomly selected, the entire generalization from five classrooms to all FCT secondary students is unsupported. If the authors can provide evidence of a true probability sample, or alternatively re-state the findings as exploratory, the conditional acceptance remains appropriate; if they cannot, the conclusions should be narrowed to the sampled schools. This does not change the reader's verdict but sharpens the condition under which the paper should be accepted.","tokens_in":21195,"tokens_out":3920,"duration_ms":47974,"concrete_test":"Ask the authors for the school-level sampling frame and selection records. If schools were not randomly drawn from a list of all FCT secondary schools, use a one-way ANOVA or mixed model to estimate the intraclass correlation (ICC) for the primary risk-exposure item across the five schools. For a school-based design with m≈82 students per school, ICC ≥ 0.02 yields a design effect of roughly 1+(m−1)ρ and an effective sample size well below 150, pushing the margin of error above ±8%. If the ICC is negligible and selection was genuinely random, the generalizability claim survives; otherwise the findings should be reframed as a descriptive case study rather than an FCT-wide estimate.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that this survey provides a comprehensive, statistically reliable picture of FCT secondary school adolescents rests on the sampling statement in Section 3.2. The paper calls the design 'stratified simple random sampling' and computes a margin of error of ±4.83% at 95% confidence from N=409 and a population of 88,721. That formula assumes a simple random sample with known selection probabilities. What is actually described is a convenience/cluster sample: five schools (two private, three public) selected by the research team, with no sampling frame, no random selection of schools, no school-level sample sizes or response rates, and acknowledged difficulty accessing private schools (Section 5.5). Students within the same school are correlated, so the effective sample size is well below 409. School-based cluster designs commonly have design effects of 2 or more for behavioral outcomes; the true margin of error may be substantially larger than ±4.83%, or not estimable at all without selection probabilities. Consequently, the descriptive percentages—including 'one in three exposed to risk'—cannot be generalized to FCT as a whole with the stated precision. The data remain a useful exploratory snapshot of five classrooms, but the headline claim of a comprehensive, representative FCT picture is not supported by the sampling evidence provided.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a mixed-methods survey of 409 senior secondary school students in Nigeria's Federal Capital Territory (FCT), organized around three research questions on digital access, online risks and coping strategies, and the roles of stakeholders in adolescent online safety. The quantitative results describe smartphone- and mobile-data-dominated access, varied frequency of internet use, a 33% overall rate of self-reported exposure to online risks (most commonly inappropriate content at 11% and scams at 9%), and a heavy reliance on blocking/reporting tools. The qualitative component analyzes 268 open-ended recommendations and identifies seven themes, including awareness and education, parental mediation, improved safety tools, stricter age restrictions, content moderation, government accountability, and resilience building. Protection Motivation Theory is used as an interpretive lens for the findings. The paper claims to provide a comprehensive, statistically reliable picture of FCT adolescents' online experiences and draws design and policy implications for parents, educators, designers, and policymakers.","tokens_in":21425,"tokens_out":3695,"duration_ms":47600,"significance":"If treated as an exploratory study, this paper makes a useful contribution to an understudied area: adolescent online safety in West Africa, and Nigeria in particular. Its strengths include centering adolescents' own voices through open-ended recommendations, providing detailed ethical procedures for school-based research with minors, and offering context-specific design implications grounded in Nigerian infrastructural and cultural realities. The use of PMT as a post-hoc interpretive framework is not circular and adds a useful structure for discussing threat and coping appraisals. However, the current presentation overstates the generalizability of the findings: the sampling design does not support the reported margin of error, and several internal numerical inconsistencies weaken the descriptive claims. With a careful reframing of the claims as exploratory and a correction of the descriptive reporting, the paper would be a valuable empirical addition to the AfriCHI and HCI4D literature.","major_comments":[{"comment":"The sampling description does not support the stated margin of error. The text calls the design 'stratified simple random sampling' and reports a ±4.83% margin of error at 95% confidence based on N=409, but the actual procedure is a convenience sample of five schools (two private, three public) with no described random selection of schools, no sampling frame, and no school-level sample sizes or response rates. Students within the same school are clustered, so the effective sample size is smaller than 409 and the simple-random-sample margin of error is not valid. This is load-bearing because the abstract and Section 1 claim a 'comprehensive' and 'statistically reliable' picture of FCT adolescents; the authors should either describe a true probability design with selection probabilities or reframe the results as an exploratory snapshot of the five participating schools and remove or heavily qualify the margin-of-error statement.","section":"§3.2, Data Collection"},{"comment":"The introduction states that 59% of respondents reported regular access to the internet 'several times a day,' but Section 4.2.2 reports that 47% (N=196) accessed the internet several times a day, while the 59% figure in Section 4.2.1 refers to mobile data as the mode of connectivity (N=241). This is not merely a typo: the introduction uses the wrong figure to summarize the central RQ1 finding, and the discrepancy should be corrected and checked against Figure 2b.","section":"§1 vs. §4.2.2"},{"comment":"The thematic counts do not add up to the stated total. The seven theme counts are N=103, 51, 42, 32, 24, 13, and 2, which sum to 267, yet the text says 268 respondents offered insights. The authors should clarify whether multiple themes per response were allowed and report the coding unit consistently; as written, the count inconsistency undermines the precision of the RQ3 qualitative summary.","section":"§4.6, Empowering Teens Through Online Safety Education and Awareness"},{"comment":"PMT is introduced after data collection as an interpretive lens, but the survey instrument contained no PMT-based measures of perceived severity, vulnerability, response efficacy, self-efficacy, or response costs. Statements such as 'this lack of perceived threat may reflect low perceived vulnerability or severity of risk' are therefore speculative interpretations rather than measured associations. The authors should explicitly label PMT as a post-hoc interpretive framework and avoid implying that the data provide evidence about the specific PMT constructs.","section":"§3.5 and §4.3"}],"minor_comments":[{"comment":"There are several typographical errors, including 'adolescnece' in Section 1, 'In more details' in Section 4.6, 'teenagersunder' in the age-restriction quote, and the possessive 'adolescents'' in the contributions list; a careful proofreading pass is needed.","section":"Throughout"},{"comment":"The text introduces 'RQ2a' but the research questions in Section 1 do not define sub-parts for RQ2; either define the sub-questions or rephrase the reference.","section":"§4.3"},{"comment":"Figures 2, 3, and 4 are referenced for important distributions, but their readability in grayscale is not assured; consider increasing label sizes and using distinguishable fills or patterns.","section":"§4.2 and §4.3 figures"},{"comment":"The sentence 'This raises critical questions about whether platform interventions are truly effective' would be stronger if the specific questions were stated explicitly, since the surrounding paragraph then supplies the relevant concerns about Western-centric moderation.","section":"§5.3"},{"comment":"Reference [7] is cited with a 2025 access date to a CyberLeninka viewer page and appears to be an older publication; the authors should verify the original source and year and cite it in the canonical form.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the AfriCHI/HCI4D scope and the empirical data are valuable, but the central generalization claim is currently unsupported by the sampling description. This is fixable by reframing the contribution as exploratory, removing or qualifying the margin-of-error statement, and correcting the internal numerical inconsistencies; no new data collection is required for these revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a valuable first empirical look at an under-studied population, and the authors deserve credit for centering Nigerian adolescents' own voices. But the paper overclaims its sampling basis: it calls a convenience sample of five FCT schools a stratified random sample and reports a ±4.83% margin of error as if the data were a probability sample. That's not supportable. The central descriptive percentages (one in three exposed to risk, blocking/reporting as top strategy) are still useful as an exploratory snapshot of these five schools, but they shouldn't be generalized to the FCT as a whole with that precision.\n\nWhat's actually new: this fills a real gap. Prior Nigerian work focused on specific risks like cyberbullying or cybercrime; this is the first combined picture of access, risk, coping, and stakeholder roles for secondary school adolescents in the FCT. The open-ended recommendations add local color that genuinely shapes design implications—NIN-based age verification, Nollywood-style resilience narratives, local-language moderation. The PMT framing is used lightly and appropriately as an interpretive lens after the fact, not fitted to the data, so there's no circularity problem.\n\nThe soft spots are real but addressable. The sampling flaw is the load-bearing one: no school-level sample sizes, no response rates, no design effect. The stress-test note is right. The authors should either report the sample as a convenience cluster sample and drop the margin of error, or provide the selection probabilities and account for clustering. Second, the introduction says 59% of respondents accessed the internet several times a day, but Section 4.2.2 reports 47% (N=196). One of those is wrong. Third, the thematic counts sum to 267 while the text says 268 responses; minor but should be fixed. Fourth, the self-report basis is acknowledged but not probed; 28% unsure about safety training is itself a finding worth more analysis. No instrument or data released, which limits reproducibility for a descriptive study.\n\nWho this is for: the AfriCHI/HCI4D and adolescent online safety communities will want this as a baseline and a source of locally grounded design ideas. It deserves serious peer review, not desk rejection, but the revision should require the sampling claims to be rewritten and the numeric inconsistencies resolved. I'd cite it for the descriptive findings with a caveat about generalizability.","headline":"A genuinely useful exploratory survey of Nigerian teens' online safety, but the sampling and margin-of-error claims overstate generalizability and need to be corrected before I'd trust the headline numbers.","tokens_in":21954,"tokens_out":2141,"would_cite":true,"duration_ms":24478,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"One in three Nigerian secondary students has faced an online risk, and most respond by blocking or reporting, a 409-student survey finds.","keywords":["Nigerian teens","adolescent online safety","Global South","culture","HCI4D","youth online safety","parental mediation","digital access"],"falsifier":"A replication using a true probability sample of FCT secondary students that includes private schools, with the same questionnaire, would settle the representativeness question: if the one-in-three risk rate and the 66% blocking/reporting rate fall outside the paper's reported ±4.8% margin of error, the sample picture fails. Comparing a subset of self-reports against actual platform logs would test whether teens' accounts of blocking, reporting, and risk exposure match their behavior.","tokens_in":21036,"feed_emoji":"🛡️","tokens_out":11647,"duration_ms":125858,"temperature":0.7,"pith_summary":"This paper reports a teen-centered survey of 409 secondary school adolescents in Nigeria's Federal Capital Territory (FCT), using paper questionnaires to map digital access, online risks, coping strategies, and support networks. It finds a moderately connected group: 59% go online several times a day, mostly through smartphones and mobile data, but shared devices and uneven connectivity shape how often and how safely they engage. About one in three reports encountering an online risk, most commonly inappropriate content (11%) and scams (9%), and blocking or reporting is the dominant response (66%), while 24% tell no one about their online problems. Interpreted through protection-motivation theory, the paper reads these patterns as the product of teens' threat perceptions and their confidence in coping options, and it uses teens' own recommendations to argue for education, parental mediation, better tools, and stronger regulation.","feed_headline":"One in three Nigerian teens report online risks","feed_subtitle":"409-student survey in Nigeria's capital: inappropriate content and scams top the list; teens say education would help.","key_machinery":"The central object is a self-administered paper questionnaire given to 409 senior secondary students in five FCT schools, two private and three public, selected through stratified simple random sampling; closed-ended items were summarized with descriptive statistics and one open-ended item was analyzed with reflexive thematic analysis. Protection motivation theory, which separates threat appraisal (perceived severity and vulnerability) from coping appraisal (response efficacy, self-efficacy, and response costs), is the interpretive lens that turns the response rates into an explanation of why teens block, report, seek help, or do nothing. The instrument produces the rates, and the lens supplies the account of what those rates mean for design and policy.","core_discovery":"The central claim is empirical: among 409 secondary school adolescents in Nigeria's Federal Capital Territory, internet access is moderate, smartphone-centered, and shaped by infrastructural limits, and about one in three has encountered an online risk, most often inappropriate content (11%) or scams (9%). The paper further claims that blocking and reporting are the dominant coping strategy (66%), parents are the main support network (36%) while only 1% consult teachers, and 24% tell no one about online problems. It also claims that online safety education is unevenly delivered: 44% report training, 28% do not, and 28% are unsure. The paper interprets these patterns through protection-motivation theory, treating safety behavior as a function of threat appraisal and coping appraisal, and derives seven recommendation themes from 268 open-ended responses, with awareness and education the most frequent.","pith_inferences":["The one-in-three risk figure may understate exposure if the 67% with no reported risk includes teens who normalized harm or lacked the vocabulary to name it; a diary or device-log study could settle this.","The paper's data on mobile data as the main gateway and shared-device use imply that low-bandwidth and SMS-based safety tools could reach more Nigerian families than app-only parental controls; this extends the paper's design direction.","Because some 12-year-olds are online despite the usual 13+ platform age minimum, age-verification rules are already being bypassed, and any national-ID-linked verification would need strong privacy safeguards before deployment.","The protection-motivation reading implies that fear-based warnings alone may not change teen behavior unless coping confidence rises; a testable extension is whether scenario-based or gamified training increases self-efficacy and reduces the 6% inaction rate."],"forward_implications":["If one in three FCT secondary students has encountered an online risk, classroom digital-safety education should treat risk exposure as a common experience rather than an exceptional one.","If blocking and reporting are the first-line response for 66% of teens, platform safety tools need to be reliable, localized, and accompanied by alternatives for teens who lack confidence to act.","If parents are the main support (36%) while teachers are consulted by only 1%, parental-mediation programs are a higher-leverage intervention than teacher training for this population.","If 28% of teens are unsure whether they ever had safety training, current education efforts are not being delivered clearly enough to register with their intended audience.","If teens' own recommendations cluster into seven themes, interventions that combine education, mediation, moderation, and regulation are more likely to fit their lived needs than any single tool."],"supporting_citations":[{"why":"Supplies the stratified simple random sampling method used to distribute the questionnaire across school strata.","marker":"[8]"},{"why":"Supplies the reflexive thematic analysis method used to derive the seven recommendation themes from open-ended responses.","marker":"[23]"},{"why":"Provides the Namibian adolescent ICT-use and risk study that this survey extends to Nigeria.","marker":"[41]"},{"why":"Provides prior Kenyan evidence of adolescent exposure to scams, stalking, and inappropriate content that motivates the study.","marker":"[54]"},{"why":"Gives the official FCT secondary-school population total used to compute the sample's margin of error.","marker":"[69]"},{"why":"Supplies the protection-motivation theory framing used to interpret threat and coping appraisals.","marker":"[74]"},{"why":"Provides prior adolescent online safety findings on parental mediation and teen self-regulation that the discussion compares against.","marker":"[79]"}],"fun_headline_variants":["1 in 3 Nigerian teens face online risks, survey finds","Nigerian teens: 1 in 3 report online risks, scams, and inappropriate content","Online risks hit 1 in 3 Nigerian teens; parents are main support","Nigerian teens' online safety: 1 in 3 face risks, most tell parents"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise, which the paper's limitations section acknowledges, is that 409 students in five FCT classrooms gave truthful, self-aware answers on a paper questionnaire and that those answers represent FCT secondary school students, including private-school students who were harder to reach.","fun_headline_variants_meta":{"raw":{"variants":["1 in 3 Nigerian teens face online risks, survey finds","Nigerian teens: 1 in 3 report online risks, scams, and inappropriate content","Online risks hit 1 in 3 Nigerian teens; parents are main support","Nigerian teens' online safety: 1 in 3 face risks, most tell parents"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000688,"raw_usage":{"total_tokens":3380,"prompt_tokens":969,"completion_tokens":2411,"prompt_tokens_details":{"cached_tokens":896},"prompt_cache_hit_tokens":896,"prompt_cache_miss_tokens":73,"completion_tokens_details":{"reasoning_tokens":2323}},"tokens_in":73,"tokens_out":2411,"duration_ms":324635,"temperature":1.0,"reasoning_tokens":2323,"cache_read_input_tokens":896,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:12:08.643143+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A replication using a true probability sample of FCT secondary students that includes private schools, with the same questionnaire, would settle the representativeness question: if the one-in-three risk rate and the 66% blocking/reporting rate fall outside the paper's reported ±4.8% margin of error, the sample picture fails. Comparing a subset of self-reports against actual platform logs would test whether teens' accounts of blocking, reporting, and risk exposure match their behavior.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the stratified simple random sampling method used to distribute the questionnaire across school strata."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the official FCT secondary-school population total used to compute the sample's margin of error."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the protection-motivation theory framing used to interpret threat and coping appraisals."},{"cited_title":"Wisniewski, Heng Xu, Mary Beth Rosson, and John M","cited_arxiv_id":null,"evidence_quote":"Provides prior adolescent online safety findings on parental mediation and teen self-regulation that the discussion compares against."}],"review_version":1}